Diss work

Schrick-Noah_CG-Ep-Modeling.tex

@@ -0,0 +1,513 @@
+\documentclass[conference]{IEEEtran}
+\RequirePackage{setspace}
+\usepackage{graphicx} % Images
+\graphicspath{ {./images/} }
+
+\usepackage{float} % Table captions on top
+\floatstyle{plaintop}
+\restylefloat{table}
+
+\usepackage{ifpdf} % Detect PDF or DVI mode
+\usepackage{babel} % Bibliography
+\usepackage{dsfont} % mathbb
+
+\usepackage[utf8]{inputenc}
+\usepackage{indentfirst}
+\setlength{\parskip}{\baselineskip}
+
+% Table of Contents/Figure Spacing
+\usepackage[titles]{tocloft} 
+\cftsetindents{figure}{0em}{3.5em}
+\cftsetindents{table}{0em}{3.5em}
+
+\usepackage{dsfont} % mathbb
+\usepackage{amsmath}
+\usepackage[linesnumbered,commentsnumbered,ruled,vlined]{algorithm2e}
+
+\begin{document}
+
+\title{ Compliance Graph Analysis Using Epidemiology Modeling
+}
+
+\author{
+    \IEEEauthorblockN{Noah L. Schrick}
+    \IEEEauthorblockA{
+        \textit{Tandy School of Computer Science} \\
+        \textit{The University of Tulsa}\\
+            Tulsa, USA \\
+            noah-schrick@utulsa.edu
+    }
+
+    \and
+
+    \IEEEauthorblockN{Peter J. Hawrylak}
+    \IEEEauthorblockA{
+        \textit{Tandy School of Computer Science} \\
+        \textit{The University of Tulsa}\\
+            Tulsa, USA \\
+            peter-hawrylak@utulsa.edu
+    }
+
+}
+
+\maketitle
+
+\begin{abstract}
+\end{abstract}
+
+\begin{IEEEkeywords}
+    Compliance Graph; Attack Graph; Epidemiology Modeling; Risk Assessment;
+\end{IEEEkeywords}
+
+\section{Introduction}
+CG Intro
+
+Epidemiology modeling is frequently used to model the behavior and dynamics of a disease across a population. Most commonly, a SIR model (Susceptible - Infected - Recovered model) is used to model how compartments of a population change with regard to a disease. The model is able to capture how the ``susceptible" subset of the population (the subset that is able to become infected) changes over time, as well as how the ``infected" subset (the subset that is currently infected with the disease) and ``recovered" subset (the subset that has been infected, but has since recovered) change over time. By examining the epidemic curves, it is possible to plan and prepare possible responses to mitigate the impact, as well as measure or predict an epidemic point. These models help provide insight on how to prepare, prevent, or mitigate challenges brought on by a disease outbreak. Epidemiology modeling has been widely investigated over recent years, and various works have been presented for modeling secondary infections through a model \cite{9914620}, additionally using multiple regression \cite{9929470} or ridge regression \cite{9628991}, and including new representations for social and community interactions \cite{9630798}. 
+
+This work discusses an epidemiology modeling approach for compliance graph analysis. This work leverages the ability of epidemiology models to measure the behavior of compliance and regulation violations in a system or set of system and determine the risk they may have. This work also leverages the predictive power of epidemiology models to present mitigation or preparation strategies for future violations. By using an epidemiology approach, it is possible to refine constraint-bound correction schemes for the predicted or expected violations that systems may face.
+
+\section{Related Works} \label{sec:ep-intro}
+The application of epidemiology modeling is not limited to modeling the dynamics of a disease. Various works have been performed to apply epidemiology modeling specifically to computer and information systems. The authors of \cite{6414589} were able to analyze the robustness of a computer network using an epidemiology model, the authors of \cite{8228672} were able to determine configurations for intrusion-detection systems, the authors of \cite{9678822} were able to model the spread of computer viruses, and the authors of \cite{5455187} were able to model the spread of malware over a network. Epidemiology modeling is also able to predict and measure social or relationship conflicts that may occur from third-party information system or information technology providers. The authors of \cite{10227113} were able to predict how these conflicts may propagate during a third-party provider implementation using an SIR model. Epidemiology models continue to see analytical improvements and usage across fields. The authors of \cite{9457692} successfully demonstrated the use of a time-series data analytics approach, and the authors of \cite{9856356} present a data-based model that is capable of integrating with, and crawling across, internet sites to minimize the manual overhead of implementing closed-form epidemiology expressions. The authors of \cite{9450250} present a wide array of techniques that work in conjunction with the fields of artificial intelligence, big data, cognition, and analytical epidemiology to showcase promising results with causal probability theory.
+
+\section{Approach to Multi-Strain Epidemiology Modeling} \label{sec:ep-modeling}
+A single-strain approach was originally used for analyzing the compliance and violation trends of environments. In a single-strain approach, all violations are grouped together as a single, generic ``violation", all warnings are grouped together as a single, generic ``warning", and all compliant systems, states, policies, and behaviors are grouped together as a single, generic ``in-good-standing" label. This approach is useful for examining the system at a high-level and for identifying concerning overall trends regarding compliance or regulation standings. However, this single-strain approach does not provide insight on the severity of violations, and does not provide distinct information on types of violations. This form of analysis is unable to discern which regulations are met or not met, and is unable to distinguish where the cause of concern may lie.  In real systems, various violations may occur, each having a different level of severity. In addition, real systems are prone to violating multiple regulations simultaneously. In order to capture and examine violations at a more granular level, a multi-strain epidemiology model will be required. The multi-strain approach functions by treating each violation as a unique strain of infection, which allows for the system to be modelled across all strains with relevant violation information properly captured.
+
+\subsection{Alternative Model Details} \label{sec:Model-Choice}
+Various works have been presented for a multi-strain epidemiology model. Specifically to computer and information systems, the authors of \cite{9593147} present a S$I_1$$I_2$SD model (Susceptible, Infected-with-Strain-1, Infected-with-Strain-2, Susceptible, Deceased) for modeling malware over Internet of Things Wireless Sensor Networks. This model is able to capture two different strains of malware that each have a dependency on the other. However, for compliance graph analysis, there can, and likely will, be a greater number of violations present. As a result, limiting an epidemiology model to two strains is not fully realistic. The authors of \cite{Lou2021} constructed a multi-strain PDE-SIS (Partial Differential Equation Susceptible - Infected - Susceptible) model that resembles the following and is capable of modeling multiple strains:\\
+
+\begin{math}
+\begin{aligned} 
+    {\left\{ \begin{array}{ll} 
+        \partial _tS=d_\mathrm{S}\Delta S +\sum _{i=1}^k\gamma _i(x)I_i -\frac{S\sum _{i=1}^k\beta _i(x)I_i}{S+\sum _{j=1}^k I_j} &{} x\in \Omega , \ t>0,\\ 
+        \partial _tI_i=d_i\Delta I_i -\gamma _i(x)I_i+\beta _i(x)\frac{SI_i}{S+\sum _{j=1}^k I_j}&{} x\in \Omega ,\ t>0, \\ 
+        d_\mathrm{S}\frac{\partial S}{\partial \vec {n}}=d_i \frac{\partial I_i}{\partial \vec {n}}=0 &{} x\in \partial \Omega ,\ t>0, \\
+        N=\int _{\Omega }[S(0,x)+\sum _{i=1}^kI_i(0,x)]\hbox {d}x>0 
+    \end{array}\right. } 
+\end{aligned}
+\end{math}
+\\
+
+Fitting compliance graph analysis to an SIS model (Susceptible, Infected, Susceptible) is achievable, but this model would lose many of the attributes, nuances, and representative details that are present in other models and real systems. Fitting each compliance graph node to only ``Susceptible" or ``Infected" compartments prevents modeling of exposure or recovery of a disease. As an alternative, the authors of \cite{10.1371/journal.pone.0257512} present a multi-strain model that would be favorable for compliance graph analysis. In this approach, the authors discuss a multi-strain SEIR (Susceptible - Exposed - Infected - Recovered) model that resembles the following:\\
+
+\begin{math}
+\begin{aligned}
+    {\left\{ \begin{array}{ll}
+    S_j(t)=P(t)-E_j(t)-I_j(t)-R_j(t),\\
+    \dot{E}_j(t)=(1-\mu(t))\beta_jS_j(t)I_j(t)-\sigma_jE_j(t),\\
+    \dot{I}_j(t)=\sigma_jE_j(t)-(\mu_j+\gamma_j)I_j(t),\\
+    \dot{R}_j(t)=\gamma_jI_j(t)-\delta_jR_j(t) 
+    \end{array}\right. }
+\end{aligned}
+\end{math}
+\\
+
+This approach includes an overall population $P$ term for determining the Susceptible compartment, mitigation measures $\mu$, recovery rate $\gamma$, and immunity loss $\delta$. This approach provides greater insight than the previous models, but still lacks potentially useful information for compliance graph analysis. In this model, recovered individuals in the system do not eventually become susceptible again, and also do not face the risk of expiring from the disease. In real systems, once an asset is compliant with a regulation, it does not become immune to violating that regulation in the future; an asset is fully able to repeatedly fall out of compliance. In addition, an alternative approach to corrective recovery (namely for legacy systems, third-party vendors or applications, or other systems that carry excess risk) is to remove the system from the environment. In epidemiology modeling, this would be represented as a system expiring or compartmentalized as ``Deceased". Though this model could provide additional parameters and compartments, it represents a solid foundational starting point for building a compatible model for compliance graph analysis, especially due to the detailed equilibrium, stability, and optimal control problem proofs. Section \ref{sec:SEIRDS-model} discusses the details of the chosen model.
+
+\subsection{SEIRDS Model} \label{sec:SEIRDS-model}
+For initial investigation and testing purposes, the selected epidemiology model for the single-strain compliance graph analysis (which has been adapted for the multi-strain approach) was a SEIRDS (Susceptible - Exposed - Infected - Recovered - Deceased) model. Unlike a SIR (Susceptible - Infected - Recovered) model, the SEIRDS model includes an exposed group to represent nodes that \textit{will} become infected, as compared to the susceptible group which \textit{can} become infected. The exposed group has additional parameters such as incubation time, which provide additional detail on when the group will transition to the infected group. This model also captures events where nodes in the infected group decease rather than recover. After recovery, there is a waning immunity period where the population can become susceptible again. A similar model can be seen in the works presented by the authors of \cite{alaya2022mathematical}. This model better fits the information present in a compliance graph than the models described in Section \ref{sec:Model-Choice}, and the contextualization of this model's compartments and parameters can be seen in Tables \ref{table:model-context} and \ref{table:model-params}. Figure \ref{fig:SEIRDS-block} shows a block model representation of the SEIRDS model. Equation \ref{eq:seirds} displays the SEIRDS representation. 
+
+\begin{table}[]
+    \scriptsize
+    \centering
+    \begin{tabular}{|c|c|c|}
+    \hline
+    \textbf{Compartment} & \textbf{Description} & \textbf{Contextualization  to Compliance Graphs}                                                                                                                        \\ \hline
+    S                    & Susceptible          & All other nodes.                                                                                                                                                        \\ \hline
+    E                    & Exposed              & \begin{tabular}[c]{@{}c@{}}Nodes flagged as at risk of compliance violation \\ by an IDS, timeframe trigger, or other.\end{tabular}                                     \\ \hline
+    I                    & Infected             & Nodes that are in violation.                                                                                                                                            \\ \hline
+    R                    & Recovered            & \begin{tabular}[c]{@{}c@{}}Nodes that were infected and were capable\\  of automatic correction e.g., certificate renewal, \\ scheduled maintenance, or other.\end{tabular} \\ \hline
+    D                    & Deceased             & \begin{tabular}[c]{@{}c@{}}Nodes removed from the compliance graph;\\ Systems that were removed from the network through \\ quarantine, DMZ, legacy removal, or other \\due to excess risk.\end{tabular}                                           \\ \hline
+    \end{tabular}
+    \caption[Compartment Descriptors for Compliance Graphs]{Compartment Descriptors for Compliance Graphs. Each compartment is contextualized to compliance graphs, with their meanings and brief explanations of how nodes may fit in compartments.}
+    \label{table:model-context}
+\end{table}
+
+\begin{table}[]
+    \scriptsize
+    \centering
+    \begin{tabular}{|c|c|c|}
+    \hline
+    \textbf{Parameter}     & \textbf{Description} & \textbf{Contextualization to Compliance Graphs}                                                                                                                       \\ \hline
+    $\beta$                & Infection Rate       & \begin{tabular}[c]{@{}c@{}}Probability of nodes falling out\\  of compliance.\end{tabular}                                                                            \\ \hline
+    $\delta$               & Incubation Period    & \begin{tabular}[c]{@{}c@{}}Once a node is at risk of falling out\\  of compliance, how long it takes to actually\\  violate a mandate or regulation.\end{tabular}     \\ \hline
+    $\gamma_{R}$           & Recovery Rate        & \begin{tabular}[c]{@{}c@{}}Probability of a system to correct\\  its violation status.\end{tabular}                                                                   \\ \hline
+    $\gamma_{D}$           & Death Rate           & \begin{tabular}[c]{@{}c@{}}Probability of removal for a system \\ in violation.\end{tabular}                                                                          \\ \hline
+    $\mu$                  & Fatality Ratio       & \begin{tabular}[c]{@{}c@{}}Natural rate at which any node may \\ be removed from the network.\end{tabular}                                                            \\ \hline
+    $\epsilon$             & Infected Import Rate & \begin{tabular}[c]{@{}c@{}}Systems that are already in violation;\\  systems that do not fall out of compliance,\\  but are already violating a mandate.\end{tabular} \\ \hline
+    $\omega$               & Waning Immunity Rate & \begin{tabular}[c]{@{}c@{}}After a system recovers, how long\\  for it to be available for violation.\end{tabular}                                                    \\ \hline
+    \end{tabular}
+    \caption[Parameter Descriptors for Compliance Graphs]{Parameter Descriptors for Compliance Graphs. Each parameter is contextualized to compliance graphs, with their meanings and brief explanations of how node relationships affect the parameters.}
+    \label{table:model-params}
+\end{table}
+
+\begin{equation}
+    \begin{gathered}
+        dS(t) = \epsilon - (\beta * S(t) * I(t)) - \mu * S(t) + \omega * R(t) \\
+        dE(t) = (\beta * S(t) * I(t)) - \delta*E(t) \\
+        dI(t) = \delta * E(t) - \gamma_{d} * I(t) - \gamma_{r} * I(t) \\
+        dR(t) = \gamma_r * I(t) - \omega * R(t) \\
+        dD(t) = \gamma_d * I(t) + \mu * S(t)
+    \label{eq:seirds}
+    \end{gathered}
+\end{equation}
+
+
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/model-flow.drawio.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[SEIRDS Epidemiology Model]{SEIRDS Epidemiology Model.  Display of each compartment in the SEIRDS epidemiology model and the mathematical representation of the relationship between compartments along with the parameters.}
+        \label{fig:SEIRDS-block}
+\end{figure}
+
+Converting the single-strain model to function as a multi-strain model requires minimal additional modeling efforts. Rather than using a single SEIRDS model for all violations, each violation is instead represented by a SEIRDS model. In an environment where there are \textit{n} violations, there are \textit{n} SEIRDS models. This is represented in Table \ref{table:multi-strain-context}, which adds an indexing term to allow for multiple violations to be represented in the SEIRDS model.
+
+\begin{table}[]
+    \scriptsize
+    \centering
+    \begin{tabular}{|c|c|c|}
+    \hline
+    \textbf{Model Variable} & \textbf{Description}                       \\ \hline
+    $S_{n}$                & Susceptible to Violation \textit{n}           \\ \hline
+    $E_{n}$                & Exposed to Violation \textit{n}               \\ \hline
+    $I_{n}$                & Infected with Violation \textit{n}            \\ \hline
+    $R_{n}$                & Recovered from Violation \textit{n}           \\ \hline
+    $D_{n}$                & Deceased from Violation \textit{n}            \\ \hline
+    $\beta_{n}$          & Infection Rate for Violation \textit{n}       \\ \hline
+    $\delta_{n}$         & Incubation Period for Violation \textit{n}    \\ \hline
+    $\gamma_{R}{n}$     & Recovery Rate for Violation \textit{n}        \\ \hline
+    $\gamma_{D}{n}$     & Death Rate for Violation \textit{n}           \\ \hline
+    $\mu_{n}$            & Fatality Ratio for Violation \textit{n}       \\ \hline
+    $\epsilon_{n}$       & Infected Import Rate for Violation \textit{n} \\ \hline
+    $\omega_{n}$         & Waning Immunity Rate for Violation \textit{n} \\ \hline  
+    \end{tabular}
+    \caption[Multi-Strain Variable Descriptors for Compliance Graphs]{Multi-Strain Variable Descriptors for Compliance Graphs. Each variable from the single-strain model can be represented with an indicator of \textit{n} to denote which SEIRDS violation model the data falls into.}
+    \label{table:multi-strain-context}
+\end{table}
+
+\subsection{Implementation Details} \label{sec:ep-model-simplify}
+Many multi-strain epidemiology models are made difficult or are constrained due to strain dependencies, competition, or extinction. In many multi-strain models, two strains $A$ and $B$ are not independent. If an individual is infected with strain $A$ or $B$, it may prevent the onset of the other strain, increase the likelihood of infection, or decrease the likelihood of infection. These dependencies often have their own dynamics, and can change over time along with the model. These strains also often compete, and strain effects can be dampened or amplified based on other strain populations. With numerous competing strains, the modeling process grows increasingly difficult or infeasible. However, a multi-strain SEIRDS model for compliance graph analysis is able to be simplified. In compliance graph analysis, a simplification can be made that each strain is fully independent - an infection (or lack thereof) of a strain does not impact any other strain. For example, for the automobile maintenance network, a node in the compliance graph can be simultaneously in need of an oil change and a headlight replacement. In addition, a strain curve can be said to have no impact on any other strain curves. Though this simplification is not fully representative and yields inaccuracy for multi-step or complex, dependent-triggered violations, it serves as a starting foundation for multi-violation analysis in a system with common violations.
+
+This work made use of both Python and R to support the epidemiology modeling. The derivation of parameters and compartmentalization of nodes for the SEIRDS model was implemented using Python and the NetworkX \cite{NetworkX} library. In addition to deriving model parameters and compartmentalizing the input compliance graph, the model preparation supports the ability to split the compliance graph into subgraphs based on time steps. Given the prior-knowledge network, the model preparation is able to identify time transitions within the graph and create subgraphs. As part of the main preparation loop, the Python function is able to take input as either the entire compliance graph or a subgraph. This preparation loop iterates through all known violations, and compartmentalizes each node according to each violation. Due to the simplification, it is possible for a node to simultaneously be classified as infected with violation $A$ while being classified as recovered from violation $B$. For each violation, each node is classified as infected if a violation edge leads to the node, exposed if a node is \textit{n}-steps away from an infected node, recovered if a recovery edge leads to the node, deceased if the node has no children, or susceptible if no other conditions match. After all compartments have been identified, parameters are obtained. The parameter definitions are presented in Equation \ref{eq:param-defs}. The compartmentalization and parameter derivation process is shown in Algorithm \ref{alg:ep-compart}.
+
+\begin{equation}
+    \begin{gathered}
+        \beta_{n} = \frac{Infected_{n} * w}{len(CG.nodes)} \\
+        \delta_{n} = 1 \\
+        \gamma_{r_{n}} = \frac{Recovered_{n} * w}{len(CG.nodes)} \\
+        \gamma_{d_{n}} = \frac{Deceased_{n} * w}{len(CG.nodes)} \\
+        \mu_{n} = \frac{Deceased_{n}}{Deceased_{n} + Recovered_{n}} \\
+        \epsilon_{n} = \frac{Imported-Infected_{n}}{len(CG.nodes)} \\
+        \omega_{n} = 1 \\
+    \label{eq:param-defs}
+    \end{gathered}
+\end{equation}
+
+For the example networks presented in Chapter \ref{ch:example-networks}, certain model parameters are statically defined. For instance, in Equation \ref{eq:param-defs} both \(\delta\) and \(\omega\) are set to 1. As Table \ref{table:multi-strain-context} specifies, \(\delta\) is the incubation period, which defines the incubation period, or how many compliance graph steps (edges) are required before a node can be classified as infected. In a compliance graph format, and specifically for the example networks presented in Chapter \ref{ch:example-networks}, there will always be exactly 1-edge delays between the labeling of an infection and the classification of an infection. Since the edges in a compliance graph are used as the transitionary elements that lead to a change in a system, for node $B$ to be classified as infected, there must be at least one edge between node $A$ and node $B$ for the infected classification to occur. This is also the case for the waning immunity rate, \(\omega\). Since there must be at least one edge to transition a node from a recovered state to a susceptible state, the waning immunity rate is also set to a static value of 1.
+
+\IncMargin{1em}
+\begin{algorithm}[htbp] \label{alg:ep-compart}
+\SetKwData{Left}{left}
+\SetKwData{This}{this}
+\SetKwData{Up}{up}
+\SetKwFunction{Union}{Union}
+\SetKwFunction{FindCompress}{FindCompress}
+\SetKwInOut{Input}{Input}\SetKwInOut{Output}{Output}
+\SetKwComment{comment}{\#}{}
+\Input{Compliance Graph, PKN, Exposure Reachability n}
+\Output{SEIRDS Model Compartments and Parameters}
+\BlankLine
+
+\textbf{STEP 1:}  Loop through Compliance Graph and Compartmentalize Nodes Based on Violation. \\
+    \For{$node$ in Compliance Graph}{
+        \For{$violation$ in possibleViolations}{
+            \If{len($node$.in-edges) == 0 and $node$ is infected with $violation$}
+                {do set importedInfection[$violation$] += 1;}
+            
+            \uIf{len($node$.out-edges) == 0 and $node$ is infected with $violation$}
+                {do set Deceased[$violation$] += 1;}   
+            \uElseIf{any $node$.in-edge has a violation identifier == $violation$}
+                {do set Infected[$violation$] += 1;}
+            \uElseIf{$node$ has $n$-step reachability to a node infected with $violation$}
+                {do set Exposed[$violation$] += 1;}
+            \uElseIf{any $node$.in-edge has a recovery identifier == $violation$}
+                {do set Recovered[$violation$] += 1;}
+            \Else 
+                {do set Susceptible[$violation$] += 1;}
+        }
+    }
+
+\textbf{STEP 2:}  Obtain Model Parameters for each Violation. \\
+    \For{$violation$ in possibleViolations}{
+        {do set $violation$[parameters] = Equation \ref{eq:param-defs}}
+    }
+
+ \caption{SEIRDS Model compartmentalization and Parameter Derivation}
+\end{algorithm}
+
+In order to better capture the recurring maintenance events, the recovery parameters can be adjusted based on a periodic function. This function can be adjusted per component and/or per mitigation strategy, and is intended to be interchanged with ease. For this work, the recovery parameters were altered to fit a cosine function. Cosine was chosen in order for the recover parameter minimum to be placed at the final known time step (t=0). A sine function with appropriate shifts could also be used, however a cosine function was used for simplicity. A few adjustments are required. To set the period of the cosine function, the default period (2$\pi$) is divided by the recurring maintenance event's time step of occurrence. Since the recovery rate must be a value between 0 and 1, a piecewise function is defined for the magnitude, where $y=max(0,cosx) \forall {x>0}$. The amplitude of the function was defined by the derived parameter from Algorithm \ref{alg:ep-compart}. This is shown in Equation \ref{eq:recovery-param}.
+
+\begin{equation}
+    \begin{gathered}
+        \gamma_{r} = \gamma_{r, derived} * \max(0,cosine(\frac{2*\pi*t}{RecurringMaintenanceTimeStep}))
+    \label{eq:recovery-param}
+    \end{gathered}
+\end{equation}
+
+For accurately capturing the pre-defined events presented in Chapter \ref{ch:example-networks}, the infection rate parameters are also designed to be dynamic. Due to the degradation of parts, health, or quality of the asset over time, the infection parameter is expected to steadily increase. The pre-defined events can be used as a baseline for increasing the infection rate to support this functionality. For instance, for the automobile maintenance network, the drive belts have a pre-defined event at time step 3 where a replacement or inspection is required. A function can be defined that increases the infection rate of the drive belt violation as the component reaches its replacement or inspection lifespan. When data is available for the lifespan of a given component, an infection rate parameter increase could model the behavior of $\lim_{t \to expectedPartLifeSpan} f(t) = 1 $. For components with pre-defined events, the ``expectedPartLifeSpan" can be obtained by identifying the event time of the part failure, repair, or replacement. For components without a pre-defined event, identifying the ``expectedPartLifeSpan" would require additional data for each component represented in the compliance graph. As previously stated, this work intends to be self-contained and independent of likelihood estimations. Though tire replacement data may be readily available, identifying the likelihood of forcibly decrypting a database (or the point of its occurrence) requires transitional estimation that is outside of this work. Instead, this work made a simplification of $ \lim_{t \to \infty} f(t) = 1 $ when the component lifespan data was unavailable.
+
+This work intended for $f(t)$ to be fully replaceable. Better cost models can replace the $f(t)$ functions to better fit to the expected changes in infection parameters. Rather than identifying unique $f(t)$ functions for each component, this work implemented two functions. When the expected component lifespan was known, Equation \ref{eq:known_ft} was used as the $f(t)$ function. When the expected component lifespan was not known, Equation \ref{eq:unknown_ft} was used as the $f(t)$ function. Equation \ref{eq:known_ft} was intended to be simplistic. Due to the exponential function, the limit approaches 1 as the component reaches its lifespan. There is a shift in the function for axis alignment, but no other modifications are implemented. This function does have a quick rise toward 1, so future adjustments to its gain may be necessary. For Equation \ref{eq:unknown_ft}, more attention was given to its rise. Since the component has an unknown life span, a gradual increase toward 1 at $t=\infty$ may be unrealistic. Likewise, a rapid gain toward 1 in the first few time steps may also be unrealistic. This function was chosen and tuned to allow for a rise toward a infection rate parameter value of 1 near 30 time steps. Since time steps are variable based on the step size selection in the network generation process, this function can be tuned to better fit specific components or networks. Figure \ref{fig:unknown-infec-rate} displays a plot of this function. After a mitigation or correction scheme is implemented for a component, the infection rate is reset to its original, derived value.
+
+\begin{equation}
+    \begin{gathered}
+        f(t) =  1-\exp({(expectedPartLifeSpan-t)})
+    \label{eq:known_ft}
+    \end{gathered}
+\end{equation}
+
+\begin{equation}
+    \begin{gathered}
+        f(t) =  1-\exp(-\frac{t}{1000}^{2.5})
+    \label{eq:unknown_ft}
+    \end{gathered}
+\end{equation}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/unknown_infec_rate.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Infection Rate Parameter as a Function of Time for Components with an Unknown Lifespan]{The infection rate parameter increases in value over time to account for degradation of parts, health, or quality of components. This Figure displays the chosen function for components with unknown lifespans. It is tuned to rise toward its limit value of 1 near 30 time steps. This function is able to be modified or replaced to better fit the degradation factor for other components or environments.}
+        \label{fig:unknown-infec-rate}
+\end{figure}
+
+
+After the Python model preparation was called with the Reticulate library \cite{reticulate} (which acts as an interface between R and Python), R code was used to evaluate the problem space using a differential equation solver. The implemented solver was ODE45, provided by the pracma library \cite{pracma}. Equation \ref{eq:seirds} was defined as a function in R, with inputs of model parameters and model compartments. This function was called through ODE45 over a specified range of time, and would produce an output of epidemic curves. Unless otherwise specified, the epidemic curves would be generated starting with time = ``0" (the last known step of the compliance graph), and would evaluate data for the next 5 years.  Since this work is approaching the compliance graph through a multi-strain lens, the differential equation solver was applied over each violation. As a result, epidemic curves were produced for each violation that provided insight on the violation trends over time. This process is displayed in Algorithm \ref{alg:ep-process}.
+
+\IncMargin{1em}
+\begin{algorithm}[htbp] \label{alg:ep-process}
+\SetKwData{Left}{left}
+\SetKwData{This}{this}
+\SetKwData{Up}{up}
+\SetKwFunction{Union}{Union}
+\SetKwFunction{FindCompress}{FindCompress}
+\SetKwInOut{Input}{Input}\SetKwInOut{Output}{Output}
+\SetKwComment{comment}{\#}{}
+\Input{Compliance Graph, PKN, Exposure Reachability n}
+\Output{Epidemic Curves}
+\BlankLine
+
+\textbf{STEP 0:} Initialize the R environment. \\
+    do initialize: \\
+        \quad Set working directory; \\
+        \quad Receive inputs; \\
+        \quad Initialize Reticulate with appropriate Python versioning and executable(s); \\
+
+\textbf{STEP 1:} Compartmentalize $Compliance Graph$ and Derive Model Parameters. \\
+    {do call Algorithm \ref{alg:ep-compart} via Reticulate;}
+
+\textbf{STEP 2:} Format Compartments and Parameters. \\
+    {do unpack $STEP \mkern5mu 1$ output and load into R variables;} \\
+    {do data type conversions and shaping across time steps and violations;} \\
+
+\textbf{STEP 3:} Produce Epidemic Curves. \\
+    \For {$violation$ in dim(S)[2]} {
+        {do set vioOut =  ODE45 with compartments=SEIRDS.compart[$violation$], \\
+            \quad parameters=SEIRDS.param[$violation$], \\
+            \quad time=[0, 5 years];} \\
+        {do plot vioOut;}
+    }
+
+ \caption{Producing Epidemic Curves for all Possible Violations in a Compliance Graph}
+\end{algorithm}
+
+\subsection{Results and Analysis} \label{sec:ep-res}
+Since a large number of epidemic curves were produced (due to the large number of possible violations across three example networks), only select epidemic curves are displayed and discussed. The selected Figures display the variations in curves based on implemented mitigation strategies. Figure \ref{fig:car-ep-oil} highlights the effectiveness of the routine maintenance for the automobile maintenance network. This Figure displays the decreases in compliance violations for oil changes due to the maintenance that occurs every 6 months or 6000 miles. Likewise, Figure \ref{fig:car-ep-drivebelt} highlights the effectiveness of the routine maintenance that takes place every 1 year and 6 months for the drive belts. Both of these Figures successfully showcase the return to compliance due to the implemented repair strategies. Figure \ref{fig:hip-ep-bad} illustrates an output for an ineffective mitigation strategy for the HIPAA network. This Figure showcases that the lack of mitigation and correction implementations leads to a rise in violations, which will be attributed to noncompliance penalties. This Figure successfully showcases the necessity for organizational attention regarding this specific violation. 
+
+Another advantage to epidemiology modeling is the ability to identify, predict, and prepare for outcomes if maintenance schedules are adjusted. Figure \ref{fig:osha_end_maint} illustrates an example for the OSHA 1910H network, where the routine maintenance schedule for a component is ended. At the last known time step (t=0), it is evident that the previous maintenance schedule was successful at correcting any known violations: the Susceptible population was at 90.9\%, the Exposed, Infected, and Deceased populations were near zero, and the Recovery population was non-zero, indicating that the maintenance schedule was successfully correcting any issues with noncompliance as they arose. Parameters were then altered to determine if maintenance schedules could be removed. For this example, the alteration included setting $\gamma_r$ to the derived value, rather than based on a function of time relative to the maintenance schedule since it was removed. Figure \ref{fig:osha_end_maint} highlights that the removal of the maintenance schedule would cause an increase in the Infected population over the next 5 years, revealing that the maintenance schedule should likely remain in place.
+
+Some Figures also displayed results with no changes in compartment trends. Figure \ref{fig:osha-ep-nothing} illustrates an epidemic curve that, though accurate, yields less interesting results. This Figure showcases the trendlines of a violation that is not seen in the given compliance graph, and there is no modeled degradation of components. Since this violation never occurs, the Susceptible population remains at 100\%, with all other compartments remaining at 0\%. However, this result provides useful insight since it demonstrates that there are no current or expected noncompliance penalties that will be incurred due to this specific violation. These Figures are able to assist in the compliance maintenance process by verifying that budgetary attention can be safely allocated to other, higher priority violations.
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/oil_change.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Epidemic Curve for the Automobile Maintenance Network's Oil Change Violation]{Epidemic Curve for the Automobile Maintenance Network's Oil Change Violation.  
+        This epidemic curve highlights the effectiveness of the routine maintenance for the automobile maintenance network, evident through the regular decreases in the Infected compartments of the SEIRDS epidemiology model. These maintenance events take place every 6 months or 6000 miles, and successfully correct the vehicle to a state of compliance with respect to the oil change violations.}
+        \label{fig:car-ep-oil}
+\end{figure}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/drive_belt.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Epidemic Curve for the Automobile Maintenance Network's Drive Belt Violation]{Epidemic Curve for the Automobile Maintenance Network's Drive Belt Violation. 
+        This epidemic curve highlights the effectiveness of the routine maintenance for the automobile maintenance network, evident through the regular decreases in the Infected compartments of the SEIRDS epidemiology model. These maintenance events take place every 1 year and 6 months, and successfully correct the vehicle to a state of compliance with respect to the drive belt violations. However, there is a steady increase of violation (the population of I) due to the increased infection rate parameter caused by the pre-defined events presented in Chapter \ref{ch:example-networks}. This Figure also displays the effectiveness at preventing the spread of a drive belt violation. If the Infected population rose to 1.0, then every node in a compliance graph or subgraph would be expected to posses that violation. Since the Infected population remains below 10\%, the mitigation strategy can be considered effective at correcting the drive belt violation before it spreads through the compliance graph or subgraph.}
+        \label{fig:car-ep-drivebelt}
+\end{figure}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/sec_asses.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Epidemic Curve for the HIPAA Network's Security Assessment Violation]{This epidemic curve highlights the ineffectiveness of the mitigation strategy for the HIPAA network, evident through the continual increase and lack of any decrease in the Infected compartments of the SEIRDS epidemiology model. There are seemingly no mitigation or correction implementations, or they are unable to successfully correct the HIPAA network to a state of compliance with respect to the security assessment violation. At the final known time step (t=0), nearly 50\% of the nodes in the last subgraph contained this violation, with the remainder within 2-step exposure of infection.}
+        \label{fig:hip-ep-bad}
+\end{figure}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/osha_end.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Epidemic Curve for the OSHA 1910H Network's  Violation]{This Figure illustrates an example for the OSHA 1910H network where the routine maintenance schedule for a component is ended. At the last known time step (t=0), it is evident that the previous maintenance schedule was successful at correcting any known violations: the Susceptible population was at 90.9\%, the Exposed, Infected, and Deceased populations were near zero, and the Recovery population was non-zero, indicating that the maintenance schedule was successfully correcting any issues with noncompliance as they arose. Parameter alters were then altered to determine the outcome if maintenance schedules could be removed. For this example, $\gamma_r$ was set to be the derived value, rather than based on a function of time relative to the maintenance schedule. This Figure highlights that the removal of the maintenance schedule causes an increase in the Infected population over the next 5 years.}
+        \label{fig:osha_end_maint}
+\end{figure}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/osha-ep-nothing.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Epidemic Curve for the OSHA 1910H Network's Inventory Violation]{This epidemic curve displays results with no changes in compartment trends for the OSHA 1910H network due to the lack of any known or expected inventory violations. This epidemic curve verifies that budgetary attention can be safely allocated to other, higher priority violations.}
+        \label{fig:osha-ep-nothing}
+\end{figure}
+
+\subsection{Validation} \label{sec:ep-valid}
+In order to validate the epidemiology modeling approach, the following characteristics were examined, and test cases were created to compare against expected behavior. The results of these tests are not included in this work, since the test results were a boolean ``pass" or ``fail". If a failed test was encountered, the validation process failed, and the methodology was flawed and in need of correction. For the work presented, each test returned a successful outcome.
+\begin{itemize}
+    \item All parameters (\(\beta\), \(\delta\), \(\gamma_{r}\), \(\gamma_{d}\), \(\mu\), \(\epsilon\), \(\omega\)) for all violations are greater than or equal to 0.0.
+    \item The \(\beta\), \(\gamma_{r}\), \(\gamma_{d}\), \(\mu\), \(\epsilon\), \(\omega\) parameters for all violations are less than or equal to 1.0.
+    \item All initial compartments (\(S\), \(E\), \(I\), \(R\), \(D\)) are greater than or equal to 0.0.
+    \item The sum of all compartments (\(S\), \(E\), \(I\), \(R\), \(D\)) for all time steps for all violations does not exceed the total number of nodes for the given time-stepped subgraph.
+    \item The sum of all compartments (\(S\), \(E\), \(I\), \(R\), \(D\)) in percentage form for all time steps for all violations is equal to 1.0.
+    \item A decrease in the \(Susceptible\) compartment causes an increase in either the \(Exposed\) or \(Infected\) compartment.
+    \item A decrease in the \(Exposed\) compartment causes an increase in the \(Infected\) compartment.
+    \item A decrease in the \(Infected\) compartment causes an increase in either the \(Recovered\) or \(Deceased\) compartment.
+    \item A decrease in the \(Recovered\) compartment causes an increase in the \(Susceptible\) compartment.
+    \item A increase in the \(Deceased\) compartment causes a decrease in the total number of nodes.
+    \item The total sum of the \(Deceased\) compartment across all time steps never decreases.
+    \item The number of nodes in the epidemiology model for the specific example networks in this work never increases.
+\end{itemize}
+
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+\section{Risk Assessment} \label{sec:risk-assessment}
+Epidemiology modeling successfully provides insight on the predicted trends and behaviors of compliance violations over time. These trends are useful for identifying the utility of any correction schemes, for future mitigation planning, or for determining which periods of time may require additional attention to prevent noncompliance. However, for presenting quantitative metrics to advisory boards, upper level staff, or management, alternative figures may yield better results, especially when requesting increases in funding. One common method for determining a quantitative metric with respect to expected or predicted challenges is risk assessment. Risk assessment allows for further insight regarding potential risks, damages, or losses, and considers any mitigations, likelihood of occurrences, and can provide information on fault-tolerance in the instance of a risk event \cite{RA}. This Section discusses the risk assessment strategies that are implemented for compliance graph analysis in Section \ref{sec:ra-imp}, with results presented in Section \ref{sec:ra-res}.
+
+\subsection{Implementation Details} \label{sec:ra-imp}
+A primary motivator for this work is to provide analysis on compliance graphs that is independent of (but compatible with) any external probability or state transition estimations. As Chapters \ref{ch:Intro}, \ref{ch:rel-works}, and \ref{ch:example-networks} discuss, though works exist for determining transitions of nodes and edges through Markov Decision Processes, empirical analysis, or investigations of the National Vulnerability Database (NVD) and Common Vulnerabilities and Exposures (CVE) list, this work aimed to stay as self-contained and self-sufficient as possible. In order to achieve this, this work presents a modification to Annualized Loss Expectancy (ALE) computations. Equation \ref{eq:ALE} displays the traditional approach for computing ALE.
+
+\begin{equation}
+    \begin{gathered}
+        ALE = Annual Rate of Occurrence * Single Loss Expectancy, \quad where: \\
+        Single Loss Expectancy = Asset Value * Exposure Factor
+    \label{eq:ALE}
+    \end{gathered}
+\end{equation}
+
+Included in the presentation of ALE in Equation \ref{eq:ALE}, Annual Rate of Occurrence (ARO) is displayed. In traditional computations of ALE, this is obtained through a likelihood estimation. Since this work is intended to stay self-contained, an alternative approach is required. The SEIRDS model presented in Section \ref{sec:SEIRDS-model} included a discussion of model parameters. Included with these parameters was a derivation of the \(\beta\) parameter in Algorithm \ref{alg:ep-compart} and Equation \ref{eq:param-defs}. As Table \ref{table:multi-strain-context} displays, the \(\beta\) parameter is used to describe the infection rate of a violation. This parameter, when presented as a yearly rate, can be directly substituted for ARO in the ALE computation. Likewise, the Noncompliance Penalty can be obtained through the prior-knowledge network, which describes the penalties for noncompliance. This is shown in Equation \ref{eq:aro-beta} as the Single Loss Expectancy. Since this work specifically does not estimate the probabilities of fines or assessments, the Exposure Factor is set to 100\%. This work aims to ensure compliance is fully met, and does not introduce any analysis that allows for organizations to base their mitigation strategies based on the likelihood of being fined. This work is binary in compliance measurement only - either the organization is maintaining a compliant standing and will not be fined, or they are in violation and will receive a penalty. 
+
+\begin{equation}
+    \begin{gathered}
+        ALE_{n} = \beta_{n} * Noncompliance Penalty
+    \label{eq:aro-beta}
+    \end{gathered}
+\end{equation}
+
+Using the modified ALE equation, a process can be defined for obtaining necessary components. The \(\beta\) parameter is obtained from the SEIRDS model parameter derivation, which receives input of a compliance graph and the prior-knowledge network. Figure \ref{fig:obj1} shows the flow diagram of computing risk assessment with ALE given a compliance graph and its input files, operating on the SEIRDS model parameters.
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/Obj1.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Risk Assessment Using Epidemiology Modeling]{Risk Assessment Using Epidemiology Modeling.  Display of the risk assessment process. Using a compliance graph and its input files, the parameters and compartments can be derived for a SEIRDS model as discussed in Section \ref{sec:ep-modeling}. Using the produced multi-strain epidemic curves, Annualized Loss Expectancy can be used to quantify the costs of compliance or regulation mandate violations.}
+        \label{fig:obj1}
+\end{figure} 
+
+\subsection{Results and Analysis} \label{sec:ra-res}
+For each example network presented in Chapter \ref{ch:example-networks}, ALE was computed for all possible violations. To display the results, bar graphs were used to highlight the magnitudes of expected losses incurred from an event of noncompliance. This display allows for a simplistic view of violations that are expected to yield unfavorable penalties, as well as those that may yield little to no penalties. The bar graphs were plotted in terms of both monetary and time penalties. Figures \ref{fig:car-ale-mon} and \ref{fig:car-ale-t} present the results for the Automobile Maintenance Network (Section \ref{sec:Automotive}) for its monetary and time expected losses, respectively. Figures \ref{fig:hip-ale-mon} and \ref{fig:hip-ale-t} present the results for the HIPAA Network (Section \ref{sec:Healthcare}) for its monetary and time expected losses, respectively. Figures \ref{fig:osha-ale-mon} and \ref{fig:osha-ale-t} present the results for the OSHA 1910H Network (Section \ref{sec:OSHA}) for its monetary and time expected losses, respectively. 
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/car-ale-mon.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Annualized Monetary Loss Expectancies for the Automobile Maintenance Network]{Annualized Monetary Loss Expectancies for the Automobile Maintenance Network. 
+        Bar graph representation of the expected monetary losses across all violations for the automobile maintenance network per year. Each loss is computed through the modified ALE equation presented in Equation \ref{eq:aro-beta}. Numerous violations have no expected losses due to the lack of presence of violation in the input compliance graph, or due to successful mitigation strategies. This is a favorable result, as it allows for provable visuals that current mitigation strategies are successful in preventing any penalties for a particular mandate, or that there is no current risk of noncompliance penalties for the possible violation.}
+        \label{fig:car-ale-mon}
+\end{figure}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/car-ale-t.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Annualized Time Loss Expectancies for the Automobile Maintenance Network]{Annualized Time Loss Expectancies for the Automobile Maintenance Network. 
+        Bar graph representation of the expected time losses across all violations for the automobile maintenance network per year. Each loss is computed through the modified ALE equation presented in Equation \ref{eq:aro-beta}. Numerous violations have no expected losses due to the lack of presence of violation in the input compliance graph, or due to successful mitigation strategies. This is a favorable result, as it allows for provable visuals that current mitigation strategies are successful in preventing any penalties for a particular mandate, or that there is no current risk of noncompliance penalties for the possible violation. However, of the possible violation occurrences, the expected losses are of low magnitude due to the availability of ride-sharing or rental cars that was built into the monetary cost model.}
+        \label{fig:car-ale-t}
+\end{figure}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/hip-ale-mon.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Annualized Monetary Loss Expectancies for the HIPAA Network]{Annualized Monetary Loss Expectancies for the HIPAA Network. 
+        Bar graph representation of the expected monetary losses across all violations for the HIPAA network per year. Each loss is computed through the modified ALE equation presented in Equation \ref{eq:aro-beta}. Numerous violations have no expected losses due to the lack of presence of violation in the input compliance graph, or due to successful mitigation strategies. This is a favorable result, as it allows for provable visuals that current mitigation strategies are successful in preventing any penalties for a particular mandate, or that there is no current risk of noncompliance penalties for the possible violation.}
+        \label{fig:hip-ale-mon}
+\end{figure}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/hip-ale-t.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Annualized Time Loss Expectancies for the HIPAA Network]{Annualized Time Loss Expectancies for the HIPAA Network. 
+        Bar graph representation of the expected time losses across all violations for the HIPAA network per year. Each loss is computed through the modified ALE equation presented in Equation \ref{eq:aro-beta}. Numerous violations have no expected losses due to the lack of presence of violation in the input compliance graph, or due to successful mitigation strategies. This is a favorable result, as it allows for provable visuals that current mitigation strategies are successful in preventing any penalties for a particular mandate, or that there is no current risk of noncompliance penalties for the possible violation. However, of the possible violation occurrences, the expected losses are of low magnitude due to the lack of closure or shutdown penalty data.}
+        \label{fig:hip-ale-t}
+\end{figure}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/osha-ale-mon.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Annualized Monetary Loss Expectancies for the OSHA 1910H Network]{Annualized Monetary Loss Expectancies for the OSHA 1910H Network. 
+        Bar graph representation of the expected monetary losses across all violations for the OSHA 1910H network per year. Each loss is computed through the modified ALE equation presented in Equation \ref{eq:aro-beta}. Numerous violations have no expected losses due to the lack of presence of violation in the input compliance graph, or due to successful mitigation strategies. This is a favorable result, as it allows for provable visuals that current mitigation strategies are successful in preventing any penalties for a particular mandate, or that there is no current risk of noncompliance penalties for the possible violation.}
+        \label{fig:osha-ale-mon}
+\end{figure}
+
+\begin{figure}[htp]
+    \includegraphics[width=\linewidth]{"./images/osha-ale-t.png"}
+    \centering
+    \vspace{.2truein} \centerline{}
+        \caption[Annualized Time Loss Expectancies for the OSHA 1910H Network]{Annualized Time Loss Expectancies for the OSHA 1910H Network. 
+        Bar graph representation of the expected time losses across all violations for the OSHA 1910H network per year. Each loss is computed through the modified ALE equation presented in Equation \ref{eq:aro-beta}. Numerous violations have no expected losses due to the lack of presence of violation in the input compliance graph, or due to successful mitigation strategies. This is a favorable result, as it allows for provable visuals that current mitigation strategies are successful in preventing any penalties for a particular mandate, or that there is no current risk of noncompliance penalties for the possible violation. However, of the possible violation occurrences, the expected losses are of low magnitude due to the lack of closure or shutdown penalty data.}
+        \label{fig:osha-ale-t}
+\end{figure}
+
+There are a few notable points of interest in each of these Figures. ALE was computed for all possible violations, and no filtering or post-processing was performed. Due to this, numerous violations list no expected losses. Though these violations were set as possibilities in the analysis methods, since there were no occurrences of noncompliance in the input compliance graph, the \(\beta\) parameter was derived to be 0, and the resulting ALE was likewise 0. This is a favorable result, as it allows for provable visuals that current mitigation strategies are successful in preventing any penalties for a particular mandate.
+
+In all Figures that presented expected time losses, magnitudes are relatively low. This is due to the input of prior-knowledge networks for each example network. For all problem spaces, identifying monetary penalties was made easier due to public availability of data. However, data describing time losses due to temporary shutdowns or closures was not readily available. Rather than estimate data, create data, or insert bias, time losses were generally not present in the prior-knowledge networks. For the Automobile Maintenance network, time losses were available due to the expected time of repair, but these losses tended to be overshadowed by monetary losses. Even in instances where repairs were lengthy, costs of public transportation or ride-sharing were built into monetary losses, rather than time losses; individuals were more likely to drop a vehicle off for repair with a return at a later date for pickup. Given more data regarding time losses, these Figures could be expected to see greater variances in magnitude.
+
+\subsection{Validation} \label{sec:ale-valid}
+In order to validate ALE, the following characteristics were examined, and test cases were created to compare against expected behavior. The results of these tests are not included in this work, since the test results were a boolean ``pass" or ``fail". If a failed test was encountered, the validation process failed, and the methodology was flawed and in need of correction. For the work presented, each test returned a successful outcome.
+\begin{itemize}
+    \item The \(\beta\) parameter for all violations is greater than or equal to 0.0.
+    \item The \(\beta\) parameter for all violations for a single time step is less than or equal to 1.0.
+    \item If the \(\beta\) parameter for any violation when substituted for ARO is greater than 1, then at least two edges are present in the compliance graph when the compliance graph is split into a subgraph across one year.
+    \item All ALE values are greater than or equal to 0.0.
+    \item If the \(\beta\) parameter for a violation is greater than or equal to 0, and is less than or equal to 1, then the resulting ALE does not exceed the costs present for that violation in the prior-knowledge network.
+    \item If the \(\beta\) parameter for a violation is greater than 1, then the resulting ALE does exceed the costs present for that violation in the prior-knowledge network.
+    \item The resulting bar graphs for displaying ALE have an equal number of bars to the number of total possible violations for the example network.
+\end{itemize}
+
+\section{Future Works}
+
+\section{Conclusions}
+This work presented and implemented a SEIRDS epidemiology model used for analyzing the trends (and expected trends) of a compliance graph. This model is able to compartmentalize a compliance graph into five unique and related sets of populations. This model also functions with a variety of tunable, customizable parameters that assist in the analysis of the spread of violations in an environment. These parameters are intended to be used alongside any user-provided data, such as the maintenance or replacement schedules or the expected lifespans of components. Each example compliance graph presented in Chapter \ref{ch:example-networks} was analyzed using this model. The results were validated as part of the validation process shown in Section \ref{sec:ep-valid}. The results presented in Section \ref{sec:ep-res} display notable visualizations about the analysis process. The methodology of this work is able to uncover valuable information regarding the compliance standing of an environment, such as the effectiveness of any mitigation, correction, or prevention strategies, and the exposure to and spread of violations. Additionally, this work allows for the testing and experimentation of mitigation strategies to provide quantifiable results as to the successes or failures of any alterations, removals, or additions toward maintaining compliance. This approach is shown to be successful in all three unique example compliance graphs, and boasts the ability to be modified or fine-tuned given additional data inputs.
+
+In addition, this work presented an approach for risk assessment through a modified Annualized Loss Expectancy (ALE) computation. This technique allows for analysis on a complex environment without the need for estimation regarding rate of occurrence that may otherwise be difficult or impossible to obtain. The results as shown in Section \ref{sec:ra-res} were presented in the form of bar graphs that display the expected losses for each resource across all examined violations. The results were validated as part of the validation process shown in Section \ref{sec:ale-valid}, and demonstrate that this approach is functional even when no noncompliance instances are observed. 
+
+
+
+
+
+\addcontentsline{toc}{section}{Bibliography}
+\bibliography{Bibliography}
+\bibliographystyle{ieeetr}
+
+\end{document}

build.sh

@@ -0,0 +1,4 @@
+#!/bin/bash
+
+bibtex ./output/Schrick-Noah_CG-Ep-Modeling
+pdflatex -output-directory=./output Schrick-Noah_CG-Ep-Modeling.tex 

output/Schrick-Noah_CG-Ep-Modeling.aux

@@ -0,0 +1,120 @@
+\relax 
+\providecommand\babel@aux[2]{}
+\@nameuse{bbl@beforestart}
+\citation{9914620}
+\citation{9929470}
+\citation{9628991}
+\citation{9630798}
+\citation{6414589}
+\citation{8228672}
+\citation{9678822}
+\citation{5455187}
+\citation{10227113}
+\citation{9457692}
+\citation{9856356}
+\citation{9450250}
+\providecommand \oddpage@label [2]{}
+\babel@aux{nil}{}
+\@writefile{toc}{\contentsline {section}{\numberline {I}Introduction}{1}{}\protected@file@percent }
+\@writefile{toc}{\contentsline {section}{\numberline {II}Related Works}{1}{}\protected@file@percent }
+\newlabel{sec:ep-intro}{{II}{1}{}{}{}}
+\@writefile{toc}{\contentsline {section}{\numberline {III}Approach to Multi-Strain Epidemiology Modeling}{1}{}\protected@file@percent }
+\newlabel{sec:ep-modeling}{{III}{1}{}{}{}}
+\citation{9593147}
+\citation{Lou2021}
+\citation{10.1371/journal.pone.0257512}
+\citation{alaya2022mathematical}
+\@writefile{toc}{\contentsline {subsection}{\numberline {\mbox  {III-A}}Alternative Model Details}{2}{}\protected@file@percent }
+\newlabel{sec:Model-Choice}{{\mbox  {III-A}}{2}{}{}{}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {\mbox  {III-B}}SEIRDS Model}{2}{}\protected@file@percent }
+\newlabel{sec:SEIRDS-model}{{\mbox  {III-B}}{2}{}{}{}}
+\citation{NetworkX}
+\@writefile{lot}{\contentsline {table}{\numberline {I}{\ignorespaces Compartment Descriptors for Compliance Graphs}}{3}{}\protected@file@percent }
+\newlabel{table:model-context}{{I}{3}{}{}{}}
+\@writefile{lot}{\contentsline {table}{\numberline {II}{\ignorespaces Parameter Descriptors for Compliance Graphs}}{3}{}\protected@file@percent }
+\newlabel{table:model-params}{{II}{3}{}{}{}}
+\newlabel{eq:seirds}{{1}{3}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {1}{\ignorespaces SEIRDS Epidemiology Model}}{3}{}\protected@file@percent }
+\newlabel{fig:SEIRDS-block}{{1}{3}{}{}{}}
+\@writefile{lot}{\contentsline {table}{\numberline {III}{\ignorespaces Multi-Strain Variable Descriptors for Compliance Graphs}}{3}{}\protected@file@percent }
+\newlabel{table:multi-strain-context}{{III}{3}{}{}{}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {\mbox  {III-C}}Implementation Details}{3}{}\protected@file@percent }
+\newlabel{sec:ep-model-simplify}{{\mbox  {III-C}}{3}{}{}{}}
+\newlabel{eq:param-defs}{{2}{4}{}{}{}}
+\newlabel{alg:ep-compart}{{1}{4}{}{}{}}
+\@writefile{loa}{\contentsline {algocf}{\numberline {1}{\ignorespaces SEIRDS Model compartmentalization and Parameter Derivation}}{4}{}\protected@file@percent }
+\newlabel{eq:recovery-param}{{3}{4}{}{}{}}
+\citation{reticulate}
+\citation{pracma}
+\newlabel{eq:known_ft}{{4}{5}{}{}{}}
+\newlabel{eq:unknown_ft}{{5}{5}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {2}{\ignorespaces Infection Rate Parameter as a Function of Time for Components with an Unknown Lifespan}}{5}{}\protected@file@percent }
+\newlabel{fig:unknown-infec-rate}{{2}{5}{}{}{}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {\mbox  {III-D}}Results and Analysis}{5}{}\protected@file@percent }
+\newlabel{sec:ep-res}{{\mbox  {III-D}}{5}{}{}{}}
+\newlabel{alg:ep-process}{{2}{6}{}{}{}}
+\@writefile{loa}{\contentsline {algocf}{\numberline {2}{\ignorespaces Producing Epidemic Curves for all Possible Violations in a Compliance Graph}}{6}{}\protected@file@percent }
+\@writefile{lof}{\contentsline {figure}{\numberline {3}{\ignorespaces Epidemic Curve for the Automobile Maintenance Network's Oil Change Violation}}{6}{}\protected@file@percent }
+\newlabel{fig:car-ep-oil}{{3}{6}{}{}{}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {\mbox  {III-E}}Validation}{6}{}\protected@file@percent }
+\newlabel{sec:ep-valid}{{\mbox  {III-E}}{6}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {4}{\ignorespaces Epidemic Curve for the Automobile Maintenance Network's Drive Belt Violation}}{7}{}\protected@file@percent }
+\newlabel{fig:car-ep-drivebelt}{{4}{7}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {5}{\ignorespaces Epidemic Curve for the HIPAA Network's Security Assessment Violation}}{7}{}\protected@file@percent }
+\newlabel{fig:hip-ep-bad}{{5}{7}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {6}{\ignorespaces Epidemic Curve for the OSHA 1910H Network's Violation}}{7}{}\protected@file@percent }
+\newlabel{fig:osha_end_maint}{{6}{7}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {7}{\ignorespaces Epidemic Curve for the OSHA 1910H Network's Inventory Violation}}{7}{}\protected@file@percent }
+\newlabel{fig:osha-ep-nothing}{{7}{7}{}{}{}}
+\citation{RA}
+\@writefile{toc}{\contentsline {section}{\numberline {IV}Risk Assessment}{8}{}\protected@file@percent }
+\newlabel{sec:risk-assessment}{{IV}{8}{}{}{}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {\mbox  {IV-A}}Implementation Details}{8}{}\protected@file@percent }
+\newlabel{sec:ra-imp}{{\mbox  {IV-A}}{8}{}{}{}}
+\newlabel{eq:ALE}{{6}{8}{}{}{}}
+\newlabel{eq:aro-beta}{{7}{8}{}{}{}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {\mbox  {IV-B}}Results and Analysis}{8}{}\protected@file@percent }
+\newlabel{sec:ra-res}{{\mbox  {IV-B}}{8}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {8}{\ignorespaces Risk Assessment Using Epidemiology Modeling}}{9}{}\protected@file@percent }
+\newlabel{fig:obj1}{{8}{9}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {9}{\ignorespaces Annualized Monetary Loss Expectancies for the Automobile Maintenance Network}}{9}{}\protected@file@percent }
+\newlabel{fig:car-ale-mon}{{9}{9}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {10}{\ignorespaces Annualized Time Loss Expectancies for the Automobile Maintenance Network}}{9}{}\protected@file@percent }
+\newlabel{fig:car-ale-t}{{10}{9}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {11}{\ignorespaces Annualized Monetary Loss Expectancies for the HIPAA Network}}{9}{}\protected@file@percent }
+\newlabel{fig:hip-ale-mon}{{11}{9}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {12}{\ignorespaces Annualized Time Loss Expectancies for the HIPAA Network}}{10}{}\protected@file@percent }
+\newlabel{fig:hip-ale-t}{{12}{10}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {13}{\ignorespaces Annualized Monetary Loss Expectancies for the OSHA 1910H Network}}{10}{}\protected@file@percent }
+\newlabel{fig:osha-ale-mon}{{13}{10}{}{}{}}
+\@writefile{lof}{\contentsline {figure}{\numberline {14}{\ignorespaces Annualized Time Loss Expectancies for the OSHA 1910H Network}}{10}{}\protected@file@percent }
+\newlabel{fig:osha-ale-t}{{14}{10}{}{}{}}
+\@writefile{toc}{\contentsline {subsection}{\numberline {\mbox  {IV-C}}Validation}{10}{}\protected@file@percent }
+\newlabel{sec:ale-valid}{{\mbox  {IV-C}}{10}{}{}{}}
+\bibdata{Bibliography}
+\bibcite{9914620}{1}
+\bibcite{9929470}{2}
+\bibcite{9628991}{3}
+\bibcite{9630798}{4}
+\bibcite{6414589}{5}
+\bibcite{8228672}{6}
+\bibcite{9678822}{7}
+\bibcite{5455187}{8}
+\bibcite{10227113}{9}
+\bibcite{9457692}{10}
+\bibcite{9856356}{11}
+\bibcite{9450250}{12}
+\bibcite{9593147}{13}
+\bibcite{Lou2021}{14}
+\bibcite{10.1371/journal.pone.0257512}{15}
+\@writefile{toc}{\contentsline {section}{\numberline {V}Future Works}{11}{}\protected@file@percent }
+\@writefile{toc}{\contentsline {section}{\numberline {VI}Conclusions}{11}{}\protected@file@percent }
+\@writefile{toc}{\contentsline {section}{Bibliography}{11}{}\protected@file@percent }
+\@writefile{toc}{\contentsline {section}{References}{11}{}\protected@file@percent }
+\bibcite{alaya2022mathematical}{16}
+\bibcite{NetworkX}{17}
+\bibcite{reticulate}{18}
+\bibcite{pracma}{19}
+\bibcite{RA}{20}
+\bibstyle{ieeetr}
+\gdef \@abspage@last{12}

output/Schrick-Noah_CG-Ep-Modeling.bbl

@@ -0,0 +1,112 @@
+\begin{thebibliography}{10}
+
+\bibitem{9914620}
+S.~Ahn and M.~Kwon, ``Reproduction factor based latent epidemic model
+  inference: A data-driven approach using covid-19 datasets,'' {\em IEEE
+  Journal of Biomedical and Health Informatics}, vol.~27, no.~3,
+  pp.~1259--1270, 2023.
+
+\bibitem{9929470}
+T.~J. Roy, M.~A. Mahmood, A.~Mohanta, and D.~Roy, ``An analytical approach to
+  predict the covid-19 death rate in bangladesh utilizing multiple regression
+  and seir model,'' in {\em 2021 IEEE International Conference on Robotics,
+  Automation, Artificial-Intelligence and Internet-of-Things (RAAICON)},
+  pp.~42--45, 2021.
+
+\bibitem{9628991}
+D.~Chumachenko, K.~Bazilevych, I.~Meniailov, S.~Yakovlev, and T.~Chumachenko,
+  ``Simulation of covid-19 dynamics using ridge regression,'' in {\em 2021 IEEE
+  4th International Conference on Advanced Information and Communication
+  Technologies (AICT)}, pp.~163--166, 2021.
+
+\bibitem{9630798}
+S.~Zhang and H.~Yang, ``Spatial modeling and analysis of human traffic and
+  infectious virus spread in community networks,'' in {\em 2021 43rd Annual
+  International Conference of the IEEE Engineering in Medicine \& Biology
+  Society (EMBC)}, pp.~2286--2289, 2021.
+
+\bibitem{6414589}
+H.~Kim and R.~Anderson, ``An experimental evaluation of robustness of
+  networks,'' {\em IEEE Systems Journal}, vol.~7, no.~2, pp.~179--188, 2013.
+
+\bibitem{8228672}
+R.~Mitchell, ``Epidemic-resistant configurations for intrusion detection
+  systems,'' in {\em 2017 IEEE Conference on Communications and Network
+  Security (CNS)}, pp.~487--494, 2017.
+
+\bibitem{9678822}
+N.~Dakhno, O.~Leshchenko, Y.~Kravchenko, A.~Dudnik, O.~Trush, and
+  V.~Khankishiev, ``Dynamic model of the spread of viruses in a computer
+  network using differential equations,'' in {\em 2021 IEEE 3rd International
+  Conference on Advanced Trends in Information Theory (ATIT)}, pp.~111--115,
+  2021.
+
+\bibitem{5455187}
+B.~Shan, ``The spread of malware on the wifi network: Epidemiology model and
+  behaviour evaluation,'' in {\em 2009 First International Conference on
+  Information Science and Engineering}, pp.~1916--1918, 2009.
+
+\bibitem{10227113}
+Y.~Tang and R.~A. Williams, ``Investigating relationship conflict within the
+  social network of large is projects using a sir model,'' in {\em 2022 IEEE
+  International Symposium on Technology and Society (ISTAS)}, vol.~1, pp.~1--5,
+  2022.
+
+\bibitem{9457692}
+M.~A. Parwez, M.~Abulaish, and J.~Jahiruddin, ``A social media time-series data
+  analytics approach for digital epidemiology,'' in {\em 2020 IEEE/WIC/ACM
+  International Joint Conference on Web Intelligence and Intelligent Agent
+  Technology (WI-IAT)}, pp.~852--859, 2020.
+
+\bibitem{9856356}
+D.~Mathebula, ``Novel data-based model for future epidemiology,'' in {\em 2022
+  International Conference on Artificial Intelligence, Big Data, Computing and
+  Data Communication Systems (icABCD)}, pp.~1--6, 2022.
+
+\bibitem{9450250}
+Y.~Wang, K.~N. Plataniotis, J.~Z. Wang, M.~Hou, M.~Zhou, N.~Howard, J.~Peng,
+  R.~Huang, S.~Patel, and D.~Zhang, ``The cognitive and mathematical
+  foundations of analytic epidemiology,'' in {\em 2020 IEEE 19th International
+  Conference on Cognitive Informatics \& Cognitive Computing (ICCI*CC)},
+  pp.~6--14, 2020.
+
+\bibitem{9593147}
+D.~Fedorov, Y.~Tabarak, A.~Dadlani, M.~S. Kumar, and V.~Kizheppatt, ``Dynamics
+  of multi-strain malware epidemics over duty-cycled wireless sensor
+  networks,'' in {\em 2021 International Balkan Conference on Communications
+  and Networking (BalkanCom)}, pp.~1--5, 2021.
+
+\bibitem{Lou2021}
+Y.~Lou and R.~B. Salako, ``Control strategies for a multi-strain epidemic
+  model,'' {\em Bulletin of Mathematical Biology}, vol.~84, p.~10, Nov 2021.
+
+\bibitem{10.1371/journal.pone.0257512}
+E.~F. Arruda, S.~S. Das, C.~M. Dias, and D.~H. Pastore, ``Modelling and optimal
+  control of multi strain epidemics, with application to covid-19,'' {\em Plos
+  One}, vol.~16, pp.~1--18, 09 2021.
+
+\bibitem{alaya2022mathematical}
+M.~B. Alaya, W.~B. Aribi, and S.~B. Miled, ``Mathematical analysis of a delayed
+  seirds epidemics models: Deterministic and stochastic approach,'' 2022.
+\newblock arXiv:2208.07690.
+
+\bibitem{NetworkX}
+A.~Hagberg, P.~J. Swart, and D.~A. Schult, ``Exploring network structure,
+  dynamics, and function using networkx,''
+\newblock {Available: https://www.osti.gov/biblio/960616}.
+
+\bibitem{reticulate}
+K.~Ushey, J.~Allaire, and Y.~Tang, {\em Reticulate: Interface to 'Python'},
+  2023.
+\newblock {R package version 1.28. Available:
+  https://CRAN.R-project.org/package=reticulate}.
+
+\bibitem{pracma}
+H.~W. Borchers, {\em Pracma: Practical Numerical Math Functions}, 2022.
+\newblock R package version 2.4.2.
+
+\bibitem{RA}
+M.~Rausand, ``Introduction,'' in {\em Risk Assessment: Theory, Methods, and
+  Applications}, ch.~1, pp.~1--28, John Wiley \& Sons Inc., 2013.
+
+\end{thebibliography}

output/Schrick-Noah_CG-Ep-Modeling.blg

@@ -0,0 +1,57 @@
+This is BibTeX, Version 0.99d (TeX Live 2024/Arch Linux)
+Capacity: max_strings=200000, hash_size=200000, hash_prime=170003
+The top-level auxiliary file: ./output/Schrick-Noah_CG-Ep-Modeling.aux
+The style file: ieeetr.bst
+Database file #1: Bibliography.bib
+You're missing a field part---line 390 of file Bibliography.bib
+ :  number = 
+ :           ,
+I'm skipping whatever remains of this entry
+Repeated entry---line 1768 of file Bibliography.bib
+ : @manual{pracma
+ :               ,
+I'm skipping whatever remains of this entry
+Warning--empty journal in NetworkX
+Warning--empty year in NetworkX
+You've used 20 entries,
+            1876 wiz_defined-function locations,
+            591 strings with 7778 characters,
+and the built_in function-call counts, 5724 in all, are:
+= -- 524
+> -- 218
+< -- 0
++ -- 81
+- -- 61
+* -- 382
+:= -- 776
+add.period$ -- 23
+call.type$ -- 20
+change.case$ -- 18
+chr.to.int$ -- 0
+cite$ -- 22
+duplicate$ -- 319
+empty$ -- 593
+format.name$ -- 61
+if$ -- 1439
+int.to.chr$ -- 0
+int.to.str$ -- 20
+missing$ -- 17
+newline$ -- 67
+num.names$ -- 20
+pop$ -- 105
+preamble$ -- 1
+purify$ -- 0
+quote$ -- 0
+skip$ -- 216
+stack$ -- 0
+substring$ -- 353
+swap$ -- 124
+text.length$ -- 0
+text.prefix$ -- 0
+top$ -- 0
+type$ -- 0
+warning$ -- 2
+while$ -- 51
+width$ -- 22
+write$ -- 189
+(There were 2 error messages)

output/Schrick-Noah_CG-Ep-Modeling.log

@@ -0,0 +1,748 @@
+This is pdfTeX, Version 3.141592653-2.6-1.40.26 (TeX Live 2024/Arch Linux) (preloaded format=pdflatex 2024.4.16)  19 APR 2024 12:00
+entering extended mode
+ restricted \write18 enabled.
+ %&-line parsing enabled.
+**Schrick-Noah_CG-Ep-Modeling.tex
+(./Schrick-Noah_CG-Ep-Modeling.tex
+LaTeX2e <2023-11-01> patch level 1
+L3 programming layer <2024-02-20>
+(/usr/share/texmf-dist/tex/latex/ieeetran/IEEEtran.cls
+Document Class: IEEEtran 2015/08/26 V1.8b by Michael Shell
+-- See the "IEEEtran_HOWTO" manual for usage information.
+-- http://www.michaelshell.org/tex/ieeetran/
+\@IEEEtrantmpdimenA=\dimen140
+\@IEEEtrantmpdimenB=\dimen141
+\@IEEEtrantmpdimenC=\dimen142
+\@IEEEtrantmpcountA=\count188
+\@IEEEtrantmpcountB=\count189
+\@IEEEtrantmpcountC=\count190
+\@IEEEtrantmptoksA=\toks17
+LaTeX Font Info:    Trying to load font information for OT1+ptm on input line 5
+03.
+(/usr/share/texmf-dist/tex/latex/psnfss/ot1ptm.fd
+File: ot1ptm.fd 2001/06/04 font definitions for OT1/ptm.
+)
+-- Using 8.5in x 11in (letter) paper.
+-- Using PDF output.
+\@IEEEnormalsizeunitybaselineskip=\dimen143
+-- This is a 10 point document.
+\CLASSINFOnormalsizebaselineskip=\dimen144
+\CLASSINFOnormalsizeunitybaselineskip=\dimen145
+\IEEEnormaljot=\dimen146
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <5> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <5> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <7> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <7> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <8> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <8> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <9> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <9> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <10> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <10> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <11> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <11> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <12> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <12> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <17> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <17> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <20> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <20> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+LaTeX Font Info:    Font shape `OT1/ptm/bx/n' in size <24> not available
+(Font)              Font shape `OT1/ptm/b/n' tried instead on input line 1090.
+LaTeX Font Info:    Font shape `OT1/ptm/bx/it' in size <24> not available
+(Font)              Font shape `OT1/ptm/b/it' tried instead on input line 1090.
+
+\IEEEquantizedlength=\dimen147
+\IEEEquantizedlengthdiff=\dimen148
+\IEEEquantizedtextheightdiff=\dimen149
+\IEEEilabelindentA=\dimen150
+\IEEEilabelindentB=\dimen151
+\IEEEilabelindent=\dimen152
+\IEEEelabelindent=\dimen153
+\IEEEdlabelindent=\dimen154
+\IEEElabelindent=\dimen155
+\IEEEiednormlabelsep=\dimen156
+\IEEEiedmathlabelsep=\dimen157
+\IEEEiedtopsep=\skip48
+\c@section=\count191
+\c@subsection=\count192
+\c@subsubsection=\count193
+\c@paragraph=\count194
+\c@IEEEsubequation=\count195
+\abovecaptionskip=\skip49
+\belowcaptionskip=\skip50
+\c@figure=\count196
+\c@table=\count197
+\@IEEEeqnnumcols=\count198
+\@IEEEeqncolcnt=\count199
+\@IEEEsubeqnnumrollback=\count266
+\@IEEEquantizeheightA=\dimen158
+\@IEEEquantizeheightB=\dimen159
+\@IEEEquantizeheightC=\dimen160
+\@IEEEquantizeprevdepth=\dimen161
+\@IEEEquantizemultiple=\count267
+\@IEEEquantizeboxA=\box51
+\@IEEEtmpitemindent=\dimen162
+\IEEEPARstartletwidth=\dimen163
+\c@IEEEbiography=\count268
+\@IEEEtranrubishbin=\box52
+) (/usr/share/texmf-dist/tex/latex/setspace/setspace.sty
+Package: setspace 2022/12/04 v6.7b set line spacing
+)
+(/usr/share/texmf-dist/tex/latex/graphics/graphicx.sty
+Package: graphicx 2021/09/16 v1.2d Enhanced LaTeX Graphics (DPC,SPQR)
+
+(/usr/share/texmf-dist/tex/latex/graphics/keyval.sty
+Package: keyval 2022/05/29 v1.15 key=value parser (DPC)
+\KV@toks@=\toks18
+)
+(/usr/share/texmf-dist/tex/latex/graphics/graphics.sty
+Package: graphics 2022/03/10 v1.4e Standard LaTeX Graphics (DPC,SPQR)
+
+(/usr/share/texmf-dist/tex/latex/graphics/trig.sty
+Package: trig 2021/08/11 v1.11 sin cos tan (DPC)
+)
+(/usr/share/texmf-dist/tex/latex/graphics-cfg/graphics.cfg
+File: graphics.cfg 2016/06/04 v1.11 sample graphics configuration
+)
+Package graphics Info: Driver file: pdftex.def on input line 107.
+
+(/usr/share/texmf-dist/tex/latex/graphics-def/pdftex.def
+File: pdftex.def 2022/09/22 v1.2b Graphics/color driver for pdftex
+))
+\Gin@req@height=\dimen164
+\Gin@req@width=\dimen165
+)
+(/usr/share/texmf-dist/tex/latex/float/float.sty
+Package: float 2001/11/08 v1.3d Float enhancements (AL)
+\c@float@type=\count269
+\float@exts=\toks19
+\float@box=\box53
+\@float@everytoks=\toks20
+\@floatcapt=\box54
+)
+\@float@every@table=\toks21
+
+(/usr/share/texmf-dist/tex/generic/iftex/ifpdf.sty
+Package: ifpdf 2019/10/25 v3.4 ifpdf legacy package. Use iftex instead.
+
+(/usr/share/texmf-dist/tex/generic/iftex/iftex.sty
+Package: iftex 2022/02/03 v1.0f TeX engine tests
+))
+(/usr/share/texmf-dist/tex/generic/babel/babel.sty
+Package: babel 2024/02/07 v24.2 The Babel package
+\babel@savecnt=\count270
+\U@D=\dimen166
+\l@unhyphenated=\language5
+
+(/usr/share/texmf-dist/tex/generic/babel/txtbabel.def)
+\bbl@readstream=\read2
+\bbl@dirlevel=\count271
+Package babel Info: You haven't specified a language as a class or package
+(babel)             option. I'll load 'nil'. Reported on input line 4258.
+
+(/usr/share/texmf-dist/tex/generic/babel/nil.ldf
+Language: nil 2024/02/07 v24.2 Nil language
+\l@nil=\language6
+))
+(/usr/share/texmf-dist/tex/latex/doublestroke/dsfont.sty
+Package: dsfont 1995/08/01 v0.1 Double stroke roman fonts
+)
+(/usr/share/texmf-dist/tex/latex/base/inputenc.sty
+Package: inputenc 2021/02/14 v1.3d Input encoding file
+\inpenc@prehook=\toks22
+\inpenc@posthook=\toks23
+)
+(/usr/share/texmf-dist/tex/latex/tools/indentfirst.sty
+Package: indentfirst 2023/07/02 v1.03 Indent first paragraph (DPC)
+)
+(/usr/share/texmf-dist/tex/latex/tocloft/tocloft.sty
+Package: tocloft 2017/08/31 v2.3i parameterised ToC, etc., typesetting
+Package tocloft Info: The document has section divisions on input line 51.
+\cftparskip=\skip51
+\cftbeforetoctitleskip=\skip52
+\cftaftertoctitleskip=\skip53
+\cftbeforesecskip=\skip54
+\cftsecindent=\skip55
+\cftsecnumwidth=\skip56
+\cftbeforesubsecskip=\skip57
+\cftsubsecindent=\skip58
+\cftsubsecnumwidth=\skip59
+\cftbeforesubsubsecskip=\skip60
+\cftsubsubsecindent=\skip61
+\cftsubsubsecnumwidth=\skip62
+\cftbeforeparaskip=\skip63
+\cftparaindent=\skip64
+\cftparanumwidth=\skip65
+\cftbeforesubparaskip=\skip66
+\cftsubparaindent=\skip67
+\cftsubparanumwidth=\skip68
+\cftbeforeloftitleskip=\skip69
+\cftafterloftitleskip=\skip70
+\cftbeforefigskip=\skip71
+\cftfigindent=\skip72
+\cftfignumwidth=\skip73
+\c@lofdepth=\count272
+\c@lotdepth=\count273
+\cftbeforelottitleskip=\skip74
+\cftafterlottitleskip=\skip75
+\cftbeforetabskip=\skip76
+\cfttabindent=\skip77
+\cfttabnumwidth=\skip78
+)
+(/usr/share/texmf-dist/tex/latex/amsmath/amsmath.sty
+Package: amsmath 2023/05/13 v2.17o AMS math features
+\@mathmargin=\skip79
+
+For additional information on amsmath, use the `?' option.
+(/usr/share/texmf-dist/tex/latex/amsmath/amstext.sty
+Package: amstext 2021/08/26 v2.01 AMS text
+
+(/usr/share/texmf-dist/tex/latex/amsmath/amsgen.sty
+File: amsgen.sty 1999/11/30 v2.0 generic functions
+\@emptytoks=\toks24
+\ex@=\dimen167
+))
+(/usr/share/texmf-dist/tex/latex/amsmath/amsbsy.sty
+Package: amsbsy 1999/11/29 v1.2d Bold Symbols
+\pmbraise@=\dimen168
+)
+(/usr/share/texmf-dist/tex/latex/amsmath/amsopn.sty
+Package: amsopn 2022/04/08 v2.04 operator names
+)
+\inf@bad=\count274
+LaTeX Info: Redefining \frac on input line 234.
+\uproot@=\count275
+\leftroot@=\count276
+LaTeX Info: Redefining \overline on input line 399.
+LaTeX Info: Redefining \colon on input line 410.
+\classnum@=\count277
+\DOTSCASE@=\count278
+LaTeX Info: Redefining \ldots on input line 496.
+LaTeX Info: Redefining \dots on input line 499.
+LaTeX Info: Redefining \cdots on input line 620.
+\Mathstrutbox@=\box55
+\strutbox@=\box56
+LaTeX Info: Redefining \big on input line 722.
+LaTeX Info: Redefining \Big on input line 723.
+LaTeX Info: Redefining \bigg on input line 724.
+LaTeX Info: Redefining \Bigg on input line 725.
+\big@size=\dimen169
+LaTeX Font Info:    Redeclaring font encoding OML on input line 743.
+LaTeX Font Info:    Redeclaring font encoding OMS on input line 744.
+\macc@depth=\count279
+LaTeX Info: Redefining \bmod on input line 905.
+LaTeX Info: Redefining \pmod on input line 910.
+LaTeX Info: Redefining \smash on input line 940.
+LaTeX Info: Redefining \relbar on input line 970.
+LaTeX Info: Redefining \Relbar on input line 971.
+\c@MaxMatrixCols=\count280
+\dotsspace@=\muskip16
+\c@parentequation=\count281
+\dspbrk@lvl=\count282
+\tag@help=\toks25
+\row@=\count283
+\column@=\count284
+\maxfields@=\count285
+\andhelp@=\toks26
+\eqnshift@=\dimen170
+\alignsep@=\dimen171
+\tagshift@=\dimen172
+\tagwidth@=\dimen173
+\totwidth@=\dimen174
+\lineht@=\dimen175
+\@envbody=\toks27
+\multlinegap=\skip80
+\multlinetaggap=\skip81
+\mathdisplay@stack=\toks28
+LaTeX Info: Redefining \[ on input line 2953.
+LaTeX Info: Redefining \] on input line 2954.
+)
+(/usr/share/texmf-dist/tex/latex/algorithm2e/algorithm2e.sty
+Package: algorithm2e 2017/07/18 v5.2 algorithms environments
+\c@AlgoLine=\count286
+
+(/usr/share/texmf-dist/tex/latex/base/ifthen.sty
+Package: ifthen 2022/04/13 v1.1d Standard LaTeX ifthen package (DPC)
+)
+\algocf@hangindent=\skip82
+
+(/usr/share/texmf-dist/tex/latex/ifoddpage/ifoddpage.sty
+Package: ifoddpage 2022/10/18 v1.2 Conditionals for odd/even page detection
+\c@checkoddpage=\count287
+)
+(/usr/share/texmf-dist/tex/latex/tools/xspace.sty
+Package: xspace 2014/10/28 v1.13 Space after command names (DPC,MH)
+)
+(/usr/share/texmf-dist/tex/latex/relsize/relsize.sty
+Package: relsize 2013/03/29 ver 4.1
+)
+\skiptotal=\skip83
+\skiplinenumber=\skip84
+\skiprule=\skip85
+\skiphlne=\skip86
+\skiptext=\skip87
+\skiplength=\skip88
+\algomargin=\skip89
+\skipalgocfslide=\skip90
+\algowidth=\dimen176
+\inoutsize=\dimen177
+\inoutindent=\dimen178
+\interspacetitleruled=\dimen179
+\interspacealgoruled=\dimen180
+\interspacetitleboxruled=\dimen181
+\algocf@ruledwidth=\skip91
+\algocf@inoutbox=\box57
+\algocf@inputbox=\box58
+\AlCapSkip=\skip92
+\AlCapHSkip=\skip93
+\algoskipindent=\skip94
+\algocf@nlbox=\box59
+\algocf@hangingbox=\box60
+\algocf@untilbox=\box61
+\algocf@skipuntil=\skip95
+\algocf@capbox=\box62
+\algocf@lcaptionbox=\skip96
+\algoheightruledefault=\skip97
+\algoheightrule=\skip98
+\algotitleheightruledefault=\skip99
+\algotitleheightrule=\skip100
+\c@algocfline=\count288
+\c@algocfproc=\count289
+\c@algocf=\count290
+\algocf@algoframe=\box63
+\algocf@algobox=\box64
+)
+(/usr/share/texmf-dist/tex/latex/l3backend/l3backend-pdftex.def
+File: l3backend-pdftex.def 2024-02-20 L3 backend support: PDF output (pdfTeX)
+\l__color_backend_stack_int=\count291
+\l__pdf_internal_box=\box65
+)
+(./output/Schrick-Noah_CG-Ep-Modeling.aux)
+\openout1 = `Schrick-Noah_CG-Ep-Modeling.aux'.
+
+LaTeX Font Info:    Checking defaults for OML/cmm/m/it on input line 27.
+LaTeX Font Info:    ... okay on input line 27.
+LaTeX Font Info:    Checking defaults for OMS/cmsy/m/n on input line 27.
+LaTeX Font Info:    ... okay on input line 27.
+LaTeX Font Info:    Checking defaults for OT1/cmr/m/n on input line 27.
+LaTeX Font Info:    ... okay on input line 27.
+LaTeX Font Info:    Checking defaults for T1/cmr/m/n on input line 27.
+LaTeX Font Info:    ... okay on input line 27.
+LaTeX Font Info:    Checking defaults for TS1/cmr/m/n on input line 27.
+LaTeX Font Info:    ... okay on input line 27.
+LaTeX Font Info:    Checking defaults for OMX/cmex/m/n on input line 27.
+LaTeX Font Info:    ... okay on input line 27.
+LaTeX Font Info:    Checking defaults for U/cmr/m/n on input line 27.
+LaTeX Font Info:    ... okay on input line 27.
+
+-- Lines per column: 56 (exact).
+(/usr/share/texmf-dist/tex/context/base/mkii/supp-pdf.mkii
+[Loading MPS to PDF converter (version 2006.09.02).]
+\scratchcounter=\count292
+\scratchdimen=\dimen182
+\scratchbox=\box66
+\nofMPsegments=\count293
+\nofMParguments=\count294
+\everyMPshowfont=\toks29
+\MPscratchCnt=\count295
+\MPscratchDim=\dimen183
+\MPnumerator=\count296
+\makeMPintoPDFobject=\count297
+\everyMPtoPDFconversion=\toks30
+) (/usr/share/texmf-dist/tex/latex/epstopdf-pkg/epstopdf-base.sty
+Package: epstopdf-base 2020-01-24 v2.11 Base part for package epstopdf
+Package epstopdf-base Info: Redefining graphics rule for `.eps' on input line 4
+85.
+
+(/usr/share/texmf-dist/tex/latex/latexconfig/epstopdf-sys.cfg
+File: epstopdf-sys.cfg 2010/07/13 v1.3 Configuration of (r)epstopdf for TeX Liv
+e
+))
+Underfull \hbox (badness 10000) in paragraph at lines 58--60
+[]\OT1/ptm/b/it/9 Index Terms\OT1/ptm/b/n/9 ---Compliance Graph; Attack Graph;
+ []
+
+[1{/var/lib/texmf/fonts/map/pdftex/updmap/pdftex.map}{/usr/share/texmf-dist/fon
+ts/enc/dvips/base/8r.enc}
+
+
+]
+Underfull \hbox (badness 1371) in paragraph at lines 76--77
+\OT1/ptm/m/n/10 systems, the authors of [[]] present a S$\OML/cmm/m/it/10 I[]$$
+I[]$\OT1/ptm/m/n/10 SD model
+ []
+
+
+Underfull \hbox (badness 2903) in paragraph at lines 76--77
+\OT1/ptm/m/n/10 (Susceptible, Infected-with-Strain-1, Infected-with-Strain-2,
+ []
+
+
+Underfull \hbox (badness 10000) in paragraph at lines 76--77
+
+ []
+
+
+Overfull \hbox (41.39592pt too wide) in paragraph at lines 78--89
+[]$[][]$ 
+ []
+
+
+Underfull \hbox (badness 10000) in paragraph at lines 78--89
+
+ []
+
+
+Underfull \hbox (badness 3118) in paragraph at lines 90--91
+[]\OT1/ptm/m/n/10 Fitting compliance graph analysis to an SIS model
+ []
+
+
+Underfull \hbox (badness 10000) in paragraph at lines 90--91
+
+ []
+
+
+Underfull \hbox (badness 10000) in paragraph at lines 92--103
+
+ []
+
+
+Underfull \vbox (badness 6493) has occurred while \output is active []
+
+
+Underfull \hbox (badness 3281) in paragraph at lines 107--108
+\OT1/ptm/m/n/10 analysis (which has been adapted for the multi-strain
+ []
+
+
+LaTeX Warning: No positions in optional float specifier.
+               Default added (so using `tbp') on input line 109.
+
+
+Overfull \hbox (16.66411pt too wide) in paragraph at lines 112--123
+ [][] 
+ []
+
+
+LaTeX Warning: No positions in optional float specifier.
+               Default added (so using `tbp') on input line 125.
+
+
+Underfull \hbox (badness 1701) in paragraph at lines 139--139
+\OT1/ptm/m/n/10 Each parameter is contextualized to compliance graphs,
+ []
+
+
+Overfull \hbox (9.88812pt too wide) in paragraph at lines 128--141
+ [][] 
+ []
+
+
+Underfull \vbox (badness 10000) has occurred while \output is active []
+
+ [2]
+<./images/model-flow.drawio.png, id=22, 728.7225pt x 281.05pt>
+File: ./images/model-flow.drawio.png Graphic file (type png)
+<use ./images/model-flow.drawio.png>
+Package pdftex.def Info: ./images/model-flow.drawio.png  used on input line 157
+.
+(pdftex.def)             Requested size: 252.0pt x 97.18964pt.
+
+
+LaTeX Warning: No positions in optional float specifier.
+               Default added (so using `tbp') on input line 166.
+
+[3 <./images/model-flow.drawio.png>]
+
+LaTeX Warning: Reference `ch:example-networks' on page 4 undefined on input lin
+e 207.
+
+
+LaTeX Warning: Reference `ch:example-networks' on page 4 undefined on input lin
+e 207.
+
+
+Underfull \vbox (badness 10000) has occurred while \output is active []
+
+
+Overfull \hbox (43.55424pt too wide) detected at line 256
+[][][]
+ []
+
+
+LaTeX Warning: Reference `ch:example-networks' on page 4 undefined on input lin
+e 258.
+
+
+Underfull \hbox (badness 2150) in paragraph at lines 258--259
+\OT1/ptm/m/n/10 repair, or replacement. For components without a pre-
+ []
+
+
+Underfull \vbox (badness 6510) has occurred while \output is active []
+
+ [4]
+Underfull \vbox (badness 10000) has occurred while \output is active []
+
+<./images/unknown_infec_rate.png, id=31, 788.19469pt x 640.64343pt>
+File: ./images/unknown_infec_rate.png Graphic file (type png)
+<use ./images/unknown_infec_rate.png>
+Package pdftex.def Info: ./images/unknown_infec_rate.png  used on input line 27
+7.
+(pdftex.def)             Requested size: 252.0pt x 204.82434pt.
+ [5 <./images/unknown_infec_rate.png>]
+Underfull \vbox (badness 5302) has occurred while \output is active []
+
+<./images/oil_change.png, id=37, 984.67876pt x 568.1225pt>
+File: ./images/oil_change.png Graphic file (type png)
+<use ./images/oil_change.png>
+Package pdftex.def Info: ./images/oil_change.png  used on input line 332.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+<./images/drive_belt.png, id=38, 984.67876pt x 568.1225pt>
+File: ./images/drive_belt.png Graphic file (type png)
+<use ./images/drive_belt.png>
+Package pdftex.def Info: ./images/drive_belt.png  used on input line 341.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+
+
+LaTeX Warning: Reference `ch:example-networks' on page 6 undefined on input lin
+e 345.
+
+
+LaTeX Warning: Reference `ch:example-networks' on page 6 undefined on input lin
+e 345.
+
+<./images/sec_asses.png, id=39, 984.67876pt x 568.1225pt>
+File: ./images/sec_asses.png Graphic file (type png)
+<use ./images/sec_asses.png>
+Package pdftex.def Info: ./images/sec_asses.png  used on input line 350.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+<./images/osha_end.png, id=40, 984.67876pt x 568.1225pt>
+File: ./images/osha_end.png Graphic file (type png)
+<use ./images/osha_end.png>
+Package pdftex.def Info: ./images/osha_end.png  used on input line 358.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+<./images/osha-ep-nothing.png, id=41, 984.67876pt x 568.1225pt>
+File: ./images/osha-ep-nothing.png Graphic file (type png)
+<use ./images/osha-ep-nothing.png>
+Package pdftex.def Info: ./images/osha-ep-nothing.png  used on input line 366.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+[6 <./images/oil_change.png>] [7 <./images/drive_belt.png> <./images/sec_asses.
+png> <./images/osha_end.png> <./images/osha-ep-nothing.png>]
+Underfull \hbox (badness 4156) in paragraph at lines 381--382
+\OT1/ptm/m/n/10 an increase in either the $\OML/cmm/m/it/10 Exposed$ \OT1/ptm/m
+/n/10 or $\OML/cmm/m/it/10 Infected$
+ []
+
+
+Underfull \hbox (badness 4001) in paragraph at lines 383--384
+[]\OT1/ptm/m/n/10 A decrease in the $\OML/cmm/m/it/10 Infected$ \OT1/ptm/m/n/10
+ compartment causes
+ []
+
+
+Underfull \hbox (badness 1527) in paragraph at lines 383--384
+\OT1/ptm/m/n/10 an increase in either the $\OML/cmm/m/it/10 Recovered$ \OT1/ptm
+/m/n/10 or $\OML/cmm/m/it/10 Deceased$
+ []
+
+
+LaTeX Warning: Reference `ch:Intro' on page 8 undefined on input line 396.
+
+
+LaTeX Warning: Reference `ch:rel-works' on page 8 undefined on input line 396.
+
+
+LaTeX Warning: Reference `ch:example-networks' on page 8 undefined on input lin
+e 396.
+
+
+Overfull \hbox (59.90427pt too wide) detected at line 404
+[][][]
+ []
+
+
+Underfull \hbox (badness 1715) in paragraph at lines 406--407
+[]\OT1/ptm/m/n/10 Included in the presentation of ALE in Equation 6[],
+ []
+
+<./images/Obj1.png, id=56, 816.04875pt x 210.7875pt>
+File: ./images/Obj1.png Graphic file (type png)
+<use ./images/Obj1.png>
+Package pdftex.def Info: ./images/Obj1.png  used on input line 418.
+(pdftex.def)             Requested size: 252.0pt x 65.09258pt.
+
+LaTeX Warning: Reference `ch:example-networks' on page 8 undefined on input lin
+e 426.
+
+
+LaTeX Warning: Reference `sec:Automotive' on page 8 undefined on input line 426
+.
+
+
+LaTeX Warning: Reference `sec:Healthcare' on page 8 undefined on input line 426
+.
+
+
+LaTeX Warning: Reference `sec:OSHA' on page 8 undefined on input line 426.
+
+[8]
+<./images/car-ale-mon.png, id=60, 984.67876pt x 568.1225pt>
+File: ./images/car-ale-mon.png Graphic file (type png)
+<use ./images/car-ale-mon.png>
+Package pdftex.def Info: ./images/car-ale-mon.png  used on input line 429.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+<./images/car-ale-t.png, id=61, 984.67876pt x 568.1225pt>
+File: ./images/car-ale-t.png Graphic file (type png)
+<use ./images/car-ale-t.png>
+Package pdftex.def Info: ./images/car-ale-t.png  used on input line 438.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+<./images/hip-ale-mon.png, id=62, 984.67876pt x 568.1225pt>
+File: ./images/hip-ale-mon.png Graphic file (type png)
+<use ./images/hip-ale-mon.png>
+Package pdftex.def Info: ./images/hip-ale-mon.png  used on input line 447.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+<./images/hip-ale-t.png, id=63, 984.67876pt x 568.1225pt>
+File: ./images/hip-ale-t.png Graphic file (type png)
+<use ./images/hip-ale-t.png>
+Package pdftex.def Info: ./images/hip-ale-t.png  used on input line 456.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+<./images/osha-ale-mon.png, id=64, 984.67876pt x 568.1225pt>
+File: ./images/osha-ale-mon.png Graphic file (type png)
+<use ./images/osha-ale-mon.png>
+Package pdftex.def Info: ./images/osha-ale-mon.png  used on input line 465.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+<./images/osha-ale-t.png, id=65, 984.67876pt x 568.1225pt>
+File: ./images/osha-ale-t.png Graphic file (type png)
+<use ./images/osha-ale-t.png>
+Package pdftex.def Info: ./images/osha-ale-t.png  used on input line 474.
+(pdftex.def)             Requested size: 252.0pt x 145.39377pt.
+ [9 <./images/Obj1.png> <./images/car-ale-mon.png> <./images/car-ale-t.png> <./
+images/hip-ale-mon.png>]
+Underfull \hbox (badness 4120) in paragraph at lines 484--485
+[]\OT1/ptm/m/n/10 In all Figures that presented expected time losses,
+ []
+
+[10 <./images/hip-ale-t.png> <./images/osha-ale-mon.png> <./images/osha-ale-t.p
+ng>]
+
+LaTeX Warning: Reference `ch:example-networks' on page 11 undefined on input li
+ne 501.
+
+
+Underfull \hbox (badness 8056) in paragraph at lines 501--502
+[]\OT1/ptm/m/n/10 This work presented and implemented a SEIRDS
+ []
+
+
+Underfull \hbox (badness 1688) in paragraph at lines 501--502
+\OT1/ptm/m/n/10 maintaining compliance. This approach is shown to be
+ []
+
+(./output/Schrick-Noah_CG-Ep-Modeling.bbl
+Underfull \hbox (badness 1681) in paragraph at lines 17--21
+\OT1/ptm/m/n/8 T. Chumachenko, ``Simulation of covid-19 dynamics using ridge
+ []
+
+
+Underfull \hbox (badness 3579) in paragraph at lines 74--78
+\OT1/ptm/m/n/8 sensor networks,'' in \OT1/ptm/m/it/8 2021 International Balkan 
+Conference on
+ []
+
+[11]
+Underfull \hbox (badness 2865) in paragraph at lines 94--97
+\OT1/ptm/m/n/8 structure, dynamics, and function using networkx,''  Available:
+ []
+
+
+Underfull \hbox (badness 6412) in paragraph at lines 99--103
+[]\OT1/ptm/m/n/8 K. Ushey, J. Allaire, and Y. Tang, \OT1/ptm/m/it/8 Reticulate:
+ Interface to
+ []
+
+)
+
+** Conference Paper **
+Before submitting the final camera ready copy, remember to:
+
+ 1. Manually equalize the lengths of two columns on the last page
+ of your paper;
+
+ 2. Ensure that any PostScript and/or PDF output post-processing
+ uses only Type 1 fonts and that every step in the generation
+ process uses the appropriate paper size.
+
+[12
+
+] (./output/Schrick-Noah_CG-Ep-Modeling.aux)
+ ***********
+LaTeX2e <2023-11-01> patch level 1
+L3 programming layer <2024-02-20>
+ ***********
+
+
+LaTeX Warning: There were undefined references.
+
+
+LaTeX Warning: Label(s) may have changed. Rerun to get cross-references right.
+
+ ) 
+Here is how much of TeX's memory you used:
+ 5655 strings out of 476076
+ 97149 string characters out of 5793775
+ 1955187 words of memory out of 5000000
+ 27705 multiletter control sequences out of 15000+600000
+ 596834 words of font info for 100 fonts, out of 8000000 for 9000
+ 14 hyphenation exceptions out of 8191
+ 57i,18n,65p,1724b,532s stack positions out of 10000i,1000n,20000p,200000b,200000s
+</usr/share/texmf-dist/fonts/type1/public/amsfonts/cm/cmex10.pfb></usr/share/
+texmf-dist/fonts/type1/public/amsfonts/cmextra/cmex7.pfb></usr/share/texmf-dist
+/fonts/type1/public/amsfonts/cm/cmmi10.pfb></usr/share/texmf-dist/fonts/type1/p
+ublic/amsfonts/cm/cmmi5.pfb></usr/share/texmf-dist/fonts/type1/public/amsfonts/
+cm/cmmi6.pfb></usr/share/texmf-dist/fonts/type1/public/amsfonts/cm/cmmi7.pfb></
+usr/share/texmf-dist/fonts/type1/public/amsfonts/cm/cmmi8.pfb></usr/share/texmf
+-dist/fonts/type1/public/amsfonts/cm/cmr10.pfb></usr/share/texmf-dist/fonts/typ
+e1/public/amsfonts/cm/cmr5.pfb></usr/share/texmf-dist/fonts/type1/public/amsfon
+ts/cm/cmr7.pfb></usr/share/texmf-dist/fonts/type1/public/amsfonts/cm/cmsy10.pfb
+></usr/share/texmf-dist/fonts/type1/public/amsfonts/cm/cmsy7.pfb></usr/share/te
+xmf-dist/fonts/type1/urw/times/utmb8a.pfb></usr/share/texmf-dist/fonts/type1/ur
+w/times/utmbi8a.pfb></usr/share/texmf-dist/fonts/type1/urw/times/utmr8a.pfb></u
+sr/share/texmf-dist/fonts/type1/urw/times/utmri8a.pfb>
+Output written on ./output/Schrick-Noah_CG-Ep-Modeling.pdf (12 pages, 392875 by
+tes).
+PDF statistics:
+ 153 PDF objects out of 1000 (max. 8388607)
+ 78 compressed objects within 1 object stream
+ 0 named destinations out of 1000 (max. 500000)
+ 71 words of extra memory for PDF output out of 10000 (max. 10000000)
+