Applied Risk Analysis for Guiding Homeland Security Policy and Decisions

SAMRAT CHATTERJEE, PhD, is Senior Operations Research/Data Scientist and the Decision Modeling and Optimization Team Lead in the Computing and Analytics Division within the National Security Directorate at Pacific Northwest National Laboratory (PNNL). He is also Affiliate Professor of Civil and Environmental Engineering with Northeastern University in Boston.ROBERT T. BRIGANTIC, PhD, is Chief Operations Research Scientist and the Statistical Modeling and Experimental Design Team Lead in the Computing and Analytics Division within the National Security Directorate at PNNL. He is also Adjunct Professor of Operations Research with the Carson College of Business at the Washington State University.ANGELA M. WATERWORTH, MS, is Senior Operations Research/Data Scientist in the Computing and Analytics Division within the National Security Directorate at PNNL.

• About the Editors xixList of Contributors xxiPreface xxvChapter Abstracts xxviiiPart I Managing National Security Risk and Policy Programs 11 On the “Influence of Scenarios to Priorities” in Risk and Security Programs 3Heimir Thorisson and James H. Lambert1.1 Introduction 31.2 Risk Programs 41.3 Canonical Questions Guiding Development of Risk Programs 61.3.1 Canonical Question I: Scope 61.3.2 Canonical Question II: Operational Design 71.3.3 Canonical Question III: Evaluation 71.4 Scenario-Based Preferences 81.5 Methodology 91.6 Demonstration of Methods 121.7 Discussion and Conclusions 20Acknowledgments 22References 222 Survey of Risk Analytic Guidelines Across the Government 25Isaac Maya, Amelia Liu, Lily Zhu, Francine Tran, Robert Creighton and CharlesWoo2.1 Department of Defense (DOD) Overview 252.1.1 Joint Risk Analysis Methodology (JRAM) for the Chairman’s Risk Assessment (CRA) 262.1.2 Mission Assurance (MA): Risk Assessment and Management for DOD Missions 292.1.3 Risk Management Guide for DOD Acquisition 312.2 Department of Justice (DOJ) 332.3 Environmental Protection Agency (EPA) Overview 362.3.1 EPA Risk Leadership 362.3.2 EPA Risk Assessment Methodology and Guidelines 372.3.3 Risk Assessment Case Studies 402.3.4 Risk Assessment Challenges of EPA 432.3.5 Review of EPA Risk Assessment/Risk Management Methodologies 432.4 National Aeronautics and Space Administration (NASA): Overview 442.4.1 NASA Risk Leadership 442.4.2 Critical Steps in NASA Risk Assessment/Risk Management 442.4.3 Risk Assessment/Risk Management Challenges of NASA 482.4.4 Review of NASA Risk Assessment/Risk Management Methodologies 492.5 Nuclear Regulatory Commission (NRC) Overview 492.5.1 NRC Leadership 512.5.2 Critical Steps in NRC Risk Assessment/Risk Management 522.5.3 Risk Assessment/Risk Management Challenges of NRC 532.5.4 Review of NRC Risk Assessment/Risk Management Methodologies 542.6 International Standards Organization (ISO) Overview 552.6.1 ISO Leadership 572.6.2 Critical Steps in ISO Risk Assessment/Risk Management 572.6.3 Risk Assessment/Risk Management Challenges of ISO 582.7 Australia Overview 582.7.1 Australia Leadership 592.7.2 Critical Steps in Australia Risk Assessment/Risk Management 602.7.3 Risk Assessment/Risk Management Challenges of Australia 612.8 UK Overview 612.8.1 UK Leadership 612.8.2 Critical Steps in UK Risk Assessment/Risk Management 622.8.3 Risk Assessment/Risk Management Challenges of the United Kingdom 65Acknowledgments 65References 653 An Overview of Risk ModelingMethods and Approaches for National Security 69Samrat Chatterjee, Robert T. Brigantic and Angela M.Waterworth3.1 Introduction 693.2 Homeland Security Risk Landscape and Missions 703.2.1 Risk Landscape 713.2.2 Security Missions 713.2.3 Risk Definitions and Interpretations from DHS Risk Lexicon 723.3 Background Review 733.3.1 1960s to 1990s: Focus on Foundational Concepts 733.3.2 The 2000s: Increased Focus on Multi-hazard Risks Including Terrorism 753.3.3 2009 to Present: Emerging Emphasis on System Resilience and Complexity 783.4 Modeling Approaches for Risk Elements 883.4.1 Threat Modeling 883.4.2 VulnerabilityModeling 883.4.2.1 Survey-Based Methods 883.4.2.2 Systems Analysis 893.4.2.3 Network-Theoretic Approaches 893.4.2.4 Structural Analysis and ReliabilityTheory 893.4.3 Consequence Modeling 893.4.3.1 Direct Impacts 893.4.3.2 Indirect Impacts 893.4.4 Risk-Informed Decision Making 903.5 Modeling Perspectives for Further Research 903.5.1 Systemic Risk and ResilienceWithin a Unified Framework 903.5.2 Characterizing Cyber and Physical Infrastructure System Behaviors and Hazards 913.5.3 Utilizing “Big” Data or Lack of Data for Generating Risk and Resilience Analytics 913.5.4 Conceptual Multi-scale, Multi-hazard Modeling Framework 923.6 Concluding Remarks 94Acknowledgments 95References 954 Comparative Risk Rankings in Support of Homeland Security Strategic Plans 101Russell Lundberg4.1 Introduction 1014.2 Conceptual Challenges in Comparative Risk Ranking 1024.3 Practical Challenges in Comparative Ranking of Homeland Security Risks 1034.3.1 Choosing a Risk Set 1044.3.1.1 Lessons from the DMRR on Hazard Set Selection 1054.3.2 Identifying Attributes to Consider 1054.3.2.1 Lessons from the DMRR on Attribute Selection 1074.3.3 Assessing Each Risk Individually 1094.3.3.1 Lessons from the DMRR on Assessing Individual Homeland Security Risks 1114.3.4 Combining Individual Risks to Develop a Comparative Risk Ranking 1124.3.4.1 Lessons from the DMRR on Comparing Homeland Security Risks 1144.4 Policy Relevance to Strategic-Level Homeland Security Risk Rankings 1164.4.1 Insights into Homeland Security Risk Rankings 1164.4.2 Risk vs. Risk Reduction 118Acknowledgments 120References 1205 A Data ScienceWorkflow for Discovering Spatial Patterns Among Terrorist Attacks and Infrastructure 125Daniel C. Fortin, Thomas Johansen, Samrat Chatterjee, GeorgeMuller and Christine Noonan5.1 Introduction 1255.2 The Data: Global Terrorism Database 1265.3 The Tools: Exploring Data Interactively Using a Custom Shiny App 1275.4 Example: Using the App to Explore ISIL Attacks 1305.5 TheModels: StatisticalModels for Terrorist Event Data 1345.6 More Data: Obtaining Regional Infrastructure Data to Build Statistical Models 1355.7 A Model: Determining the Significance of Infrastructure on the Likelihood of an Attack 1375.8 Case Study: Libya 1385.9 Case Study: Jammu and Kashmir Region of India 1395.9.1 The Model Revisited: Accounting for Many Regions with No Recorded Attacks 1415.9.2 Investigating the Effect of Outliers 1455.9.3 The Insight: What Have We Learned? 1475.10 Summary 148References 148Part II Strengthening Ports of Entry 1516 Effects of Credibility of Retaliation Threats in Deterring Smuggling of Nuclear Weapons 153Xiaojun Shan and Jun Zhuang6.1 Introduction 1536.2 Extending Prior Game-Based Model 1586.3 Comparing the Game Trees 1586.4 The Extended Model 1616.5 Solution to the Extended Model 1626.6 Comparing the Solutions in Prior Game-Based Model and This Study 1636.7 Illustration of the Extended Model Using Real Data 1646.8 Conclusion and Future Research Work 165References 1677 Disutility of Mass Relocation After a Severe Nuclear Accident 171VickiM. Bier and Shuji Liu7.1 Introduction 1717.2 Raw Data 1747.3 Trade-Offs Between Cancer Fatalities and Relocation 1777.4 Risk-Neutral DisutilityModel 1797.5 Risk-Averse DisutilityModel 1797.6 DisutilityModel with Interaction Effects 1827.7 Economic Analysis 1857.8 Conclusion 190References 1918 Scheduling Federal Air Marshals Under Uncertainty 193KeithW. DeGregory and Rajesh Ganesan8.1 Introduction 1938.2 Literature 1968.2.1 Commercial Aviation Industry 1968.2.2 Homeland Security and the Federal Air Marshals Service 1988.2.3 Approximate Dynamic Programming 1998.3 Air Marshal Resource Allocation Model 2008.3.1 Risk Model 2008.3.2 Static Allocation 2028.3.3 Dynamic Allocation 2038.4 Stochastic Dynamic Programming Formulation 2048.4.1 System State 2058.4.2 Decision Variable 2058.4.3 Post-decision State 2068.4.4 Exogenous Information 2068.4.5 State Transition Function 2068.4.6 Contribution Function 2068.4.7 Objective Function 2078.4.8 Bellman’s Optimality Equations 2078.5 Phases of Stochastic Dynamic Programming 2078.5.1 Exploration Phase 2078.5.2 Learning Phase 2088.5.2.1 Algorithm 2088.5.2.2 Approximation Methods 2088.5.2.3 Convergence 2098.5.3 Learned Phase 2108.6 Integrated Allocation Model 2108.7 Results 2118.7.1 Experiment 2118.7.2 Results from Stochastic Dynamic Programming Model 2118.7.3 Sensitivity Analysis 2128.7.4 Model Output 2148.8 Conclusion 217Acknowledgments 218References 218Part III Securing Critical Cyber Assets 2219 Decision Theory for Network Security: Active Sensing for Detection and Prevention of Data Exfiltration 223Sara M. McCarthy, Arunesh Sinha,Milind Tambe and Pratyusa Manadhatha9.1 Introduction 2239.1.1 Problem Domain 2249.2 Background and RelatedWork 2269.2.1 DNS Exfiltration 2269.2.2 Partially Observable Markov Decision Process (POMDP) 2289.3 Threat Model 2299.3.1 The POMDP Model 2309.4 POMDP Abstraction 2329.4.1 Abstract Actions 2329.4.2 Abstract Observations 2349.4.3 VD-POMDP Factored Representation 2349.4.4 Policy Execution 2369.5 VD-POMDP Framework 2399.6 Evaluation 2419.6.1 Synthetic Networks 2419.6.2 DETER Testbed Simulation 2419.6.3 Runtime 2429.6.4 Performance 2449.6.5 Robustness 2469.7 GameTheoretic Extensions 2479.7.1 Threat Model 2489.8 Conclusion and FutureWork 249Acknowledgments 249References 24910 Measurement of Cyber Resilience from an Economic Perspective 253Adam Z. Rose and NoahMiller10.1 Introduction 25310.2 Economic Resilience 25410.2.1 Basic Concepts of Cyber Resilience 25410.2.2 Basic Concepts of Economic Resilience 25410.2.3 Economic Resilience Metrics 25510.3 Cyber System Resilience Tactics 25710.4 Resilience for Cyber-Related Sectors 26710.4.1 Resilience in the Manufacturing of Cyber Equipment 26710.4.2 Resilience in the Electricity Sector 26810.5 Conclusion 269References 27011 Responses to Cyber Near-Misses: A Scale to Measure Individual Differences 275Jinshu Cui, Heather Rosoff and Richard S. John11.1 Introduction 27511.2 Scale Development and Analysis Outline 27711.3 Method 27811.3.1 Measures 27811.3.1.1 Cyber Near-Miss Appraisal Scale (CNMAS) 27811.3.1.2 Measures of Discriminant Validity 28111.3.1.3 Measure of Predictive Validity 28111.3.1.4 Participants and Procedures 28111.4 Results 28411.4.1 Dimensionality and Reliability 28411.4.2 Item Response Analysis 28411.4.3 Differential Item Functioning (DIF) 28711.4.4 Effects of Demographic Variables 28911.4.5 Discriminant Validity 29011.4.6 Predictive Validity 29011.5 Discussion 291Acknowledgments 292References 292Part IV Enhancing Disaster Preparedness and Infrastructure Resilience 29512 An InteractiveWeb-Based Decision Support Systemfor Mass Dispensing, Emergency Preparedness, and Biosurveillance 297Eva K. Lee, Ferdinand H. Pietz, Chien-Hung Chen and Yifan Liu12.1 Introduction 29712.2 System Architecture and Design 29912.3 System Modules and Functionalities 30112.3.1 Interactive User Experience 30112.3.2 Geographical Boundaries 30112.3.3 Network of Service, Locations, and Population Flow and Assignment 30212.3.4 ZIP Code and Population Composition 30412.3.5 Multimodality Dispensing and Public–Private Partnership 30512.3.6 POD Layout Design and Resource Allocation 30812.3.7 Radiological Module 30912.3.8 Biosurveillance 30912.3.9 Regional Information Sharing, Reverse Reporting, Tracking and Monitoring, and Resupply 31012.3.10 Multilevel End-User Access 31112.4 Biodefense, Pandemic Preparedness Planning, and Radiological and Large-Scale Disaster Relief Efforts 31212.4.1 Biodefense Mass Dispensing Regional Planning 31212.4.2 Real-Life Disaster Response Effort 31512.4.2.1 RealOpt-Haiti© 31512.4.2.2 RealOpt-Regional and RealOpt-CRC for Fukushima Daiichi Nuclear Disaster 31612.4.2.3 RealOpt-ASSURE© 31812.5 Challenges and Conclusions 319Acknowledgments 321References 32113 Measuring Critical Infrastructure Risk, Protection, and Resilience in an All-Hazards Environment 325Julia A. Phillips and Frédéric Petit13.1 Introduction to Critical Infrastructure Risk Assessment 32513.2 Motivation for Critical Infrastructure Risk Assessments 32613.2.1 Unrest pre-September 2001 32613.2.2 Post-911 Critical Infrastructure Protection and Resilience 32613.3 Decision Analysis Methodologies for Creating Critical Infrastructure Risk Indicators 32713.3.1 Decision Analysis 32813.3.2 Illustrative Calculations for an Index: Buying a Car 32813.4 An Application of Critical Infrastructure Protection, Consequence, and Resilience Assessment 33113.4.1 Protection and Vulnerability 33413.4.1.1 Physical Security 33513.4.1.2 Security Management 33513.4.1.3 Security Force 33513.4.1.4 Information Sharing 33713.4.1.5 Security Activity Background 33813.4.2 Resilience 33913.4.2.1 Preparedness 34113.4.2.2 Mitigation Measures 34113.4.2.3 Response Capabilities 34213.4.2.4 Recovery Mechanisms 34313.4.3 Consequences 34313.4.3.1 Human Consequences 34513.4.3.2 Economic Consequences 34613.4.3.3 Government Mission/Public Health/Psychological Consequences 34613.4.3.4 Cascading Impact Consequences 34713.4.4 Risk Indices Comparison 34913.5 Infrastructure Interdependencies 35013.6 What’s Next for Critical Infrastructure Risk Assessments 352References 35414 Risk AnalysisMethods in Resilience Modeling: An Overview of Critical Infrastructure Applications 357Hiba Baroud14.1 Introduction 35714.2 Background 35814.2.1 Risk Analysis 35814.2.2 Resilience 35914.2.3 Critical Infrastructure Systems 36014.3 Modeling the Resilience of Critical Infrastructure Systems 36114.3.1 Resilience Models 36114.3.1.1 Manufacturing 36114.3.1.2 Communications 36214.3.1.3 Dams, Levees, andWaterways 36314.3.1.4 Defense 36314.3.1.5 Emergency Services 36314.3.1.6 Energy 36314.3.1.7 Transportation 36414.3.1.8 Water/Wastewater 36414.3.2 Discussion 36514.3.2.1 Economic Impact 36514.3.2.2 Social Impact 36714.3.2.3 Interdependencies 36714.4 Assessing Risk in Resilience Models 36814.4.1 Probabilistic Methods 36814.4.2 UncertaintyModeling 36914.4.3 Simulation-Based Approaches 36914.4.4 Data-Driven Analytics 37014.5 Opportunities and Challenges 37014.5.1 Opportunities 37014.5.2 Challenges 37114.6 Concluding Remarks 372References 37315 Optimal Resource Allocation Model to Prevent, Prepare, and Respond to Multiple Disruptions, with Application to the Deepwater Horizon Oil Spill and Hurricane Katrina 381Cameron A.MacKenzie and Amro Al Kazimi15.1 Introduction 38115.2 Model Development 38315.2.1 Resource Allocation Model 38315.2.2 Extension to Uncertain Parameters 38515.3 Application: Deepwater Horizon and Hurricane Katrina 38615.3.1 Parameter Estimation 38615.3.1.1 Oil Spill Parameters 38715.3.1.2 Hurricane Parameters 38815.3.2 Base Case Results 39115.3.3 Sensitivity Analysis on Economic Impacts 39415.3.4 Model with Uncertain Effectiveness 39515.4 Conclusions 397References 39816 Inoperability Input–Output Modeling of Electric Power Disruptions 405Joost R. Santos, Sheree Ann Pagsuyoin and Christian Yip16.1 Introduction 40516.2 Risk Analysis of Natural and Man-Caused Electric Power Disruptions 40716.3 Risk Management Insights for Disruptive Events 40816.4 Modeling the Ripple Effects for Disruptive Events 41116.5 Inoperability Input–Output Model 41216.5.1 Model Parameters 41216.5.2 Sector Inoperability 41316.5.3 InterdependencyMatrix 41316.5.4 Demand Perturbation 41416.5.5 Economic Resilience 41416.5.6 Economic Loss 41516.6 Sample Electric Power Disruptions Scenario Analysis for the United States 41616.7 Summary and Conclusions 421References 42217 Quantitative Assessment of Transportation Network Vulnerability with Dynamic Traffic Simulation Methods 427Venkateswaran Shekar and Lance Fiondella17.1 Introduction 42717.2 Dynamic Transportation Network Vulnerability Assessment 42917.3 Sources of Input for Dynamic Transportation Network Vulnerability Assessment 43117.4 Illustrations 43217.4.1 Example 1: Simple Network 43217.4.2 Example II: University of Massachusetts Dartmouth Evacuation 43717.5 Conclusion and Future Research 439References 44018 Infrastructure Monitoring for Health and Security 443Prodyot K. Basu18.1 Introduction 44318.2 Data Acquisition 44718.3 Sensors 44718.3.1 Underlying Principles of Some of the Popular Sensors Listed in Table 18.1 45118.3.1.1 Fiber Optics 45118.3.1.2 VibratingWire 45118.3.1.3 Piezoelectric Sensors 45618.3.1.4 Piezoresistive Sensors 45618.3.1.5 Laser Vibrometer 45618.3.1.6 Acoustic Emission Sensing 45718.3.1.7 GPS and GNSS 45818.3.2 Selection of a Sensor 45918.4 Capturing and Transmitting Signals 45918.5 Energy Harvesting 46118.6 Robotic IHM 46218.7 Cyber-Physical Systems 46418.8 Conclusions 464References 46519 Exploring Metaheuristic Approaches for Solving the Traveling Salesman Problem Applied to Emergency Planning and Response 467Ramakrishna Tipireddy, Javier Rubio-Herrero, Samrat Chatterjee and Satish Chikkagoudar19.1 The Traveling Salesman Problem 46719.1.1 Definition 46719.1.2 Computational Complexity 46719.1.3 Solution Algorithms 46819.1.4 Emergency Response Application 46819.2 Emergency Planning and Response as a Traveling Salesman Problem 46819.3 Metaheuristic Approaches 46919.3.1 Simulated Annealing 47019.3.1.1 Overview 47019.3.1.2 Pseudocode 47119.3.1.3 Case Study Results 47319.3.2 Tabu Search 47319.3.2.1 Overview 47319.3.2.2 Pseudocode 47419.3.2.3 Case Study Results 47619.3.3 Genetic Algorithms 47619.3.3.1 Overview 47619.3.3.2 Pseudocode 47819.3.3.3 Case Study Results 47919.3.4 Ant Colony Optimization 47919.3.4.1 Overview 47919.3.4.2 Stochastic Solution Construction 48019.3.4.3 Pheromone Update 48019.3.4.4 Pseudocode 48119.3.4.5 Case Study Results 48119.4 Discussion 48219.5 Concluding Remarks 482References 484Index 487