Two-stage invest–defend game: Balancing strategic and operational decisions

Abdolmajid Yolmeh, Melike Baykal-Gürsoy

Research output: Contribution to journalArticle

Abstract

Protecting infrastructures and their users against terrorist attacks involves making both strategic and operational decisions in an organization’s hierarchy. Although usually analyzed separately, these decisions influence each other. To study the combined effect of strategic and operational decisions, we present a game-theoretic, two-stage model between a defender and an attacker involving multiple target sites. In the first stage, the defender (attacker) makes a strategic decision of allocating investment resources to target sites to improve the defense (attack) capabilities. We consider two cases for investments in the first stage: (1) unconstrained and (2) budget-constrained models. The investment allocations for each target site determine its detection probability. In the second stage, the players make operational decisions of which target site to defend or to attack. We distinguish between two games that arise in the second stage: the maximal damage game and the infiltration/harassment game. We prove that the solution to each game under budget constraints is unique. In fact, when the second-stage game is of the infiltration/harassment type, the invest–defend game has a unique closed-form solution that is very intuitive. The results reveal that an increase in defense investments at a target site decreases the probability of both defending and attacking that target. However, an increase in attack investments increases the probability of both defending and attacking that target. Similarly, an increase in the defender’s (attacker’s) investment efficiency leads to a decrease (increase) in investments of both the defender and the attacker. Finally, the model is applied to real data to obtain the equilibrium investment and defense strategies. The results from real data demonstrate that the attacker’s penalty from a failed attack is an important factor in determining the defender’s optimal distribution of investments and defense probabilities. The defender’s second-stage defense decisions complement the first-stage investment decisions; that is, among target sites that receive little or zero investment, the most important one is covered with a relatively high defense probability in the second stage. Moreover, as the attacker’s budget increases, the defense investments shift from less important sites to more important ones.

Original languageEnglish (US)
Pages (from-to)46-66
Number of pages21
JournalDecision Analysis
Volume16
Issue number1
DOIs
StatePublished - Mar 2019

Fingerprint

Two-stage game
Attack
Harassment
Investment allocation
Two-stage model
Penalty
Budget constraint
Damage
Factors
Closed-form solution
Investment decision
Strategic decisions
Resources
Investment efficiency
Terrorist attack

All Science Journal Classification (ASJC) codes

  • Decision Sciences(all)

Keywords

  • Budget constraint
  • Infrastructure security games
  • Sequential game
  • Simultaneous game
  • Strategic decision making
  • Terrorism and counterterrorism
  • Two-stage game

Cite this

@article{270105153958497bad39c440a108a132,
title = "Two-stage invest–defend game: Balancing strategic and operational decisions",
abstract = "Protecting infrastructures and their users against terrorist attacks involves making both strategic and operational decisions in an organization’s hierarchy. Although usually analyzed separately, these decisions influence each other. To study the combined effect of strategic and operational decisions, we present a game-theoretic, two-stage model between a defender and an attacker involving multiple target sites. In the first stage, the defender (attacker) makes a strategic decision of allocating investment resources to target sites to improve the defense (attack) capabilities. We consider two cases for investments in the first stage: (1) unconstrained and (2) budget-constrained models. The investment allocations for each target site determine its detection probability. In the second stage, the players make operational decisions of which target site to defend or to attack. We distinguish between two games that arise in the second stage: the maximal damage game and the infiltration/harassment game. We prove that the solution to each game under budget constraints is unique. In fact, when the second-stage game is of the infiltration/harassment type, the invest–defend game has a unique closed-form solution that is very intuitive. The results reveal that an increase in defense investments at a target site decreases the probability of both defending and attacking that target. However, an increase in attack investments increases the probability of both defending and attacking that target. Similarly, an increase in the defender’s (attacker’s) investment efficiency leads to a decrease (increase) in investments of both the defender and the attacker. Finally, the model is applied to real data to obtain the equilibrium investment and defense strategies. The results from real data demonstrate that the attacker’s penalty from a failed attack is an important factor in determining the defender’s optimal distribution of investments and defense probabilities. The defender’s second-stage defense decisions complement the first-stage investment decisions; that is, among target sites that receive little or zero investment, the most important one is covered with a relatively high defense probability in the second stage. Moreover, as the attacker’s budget increases, the defense investments shift from less important sites to more important ones.",
keywords = "Budget constraint, Infrastructure security games, Sequential game, Simultaneous game, Strategic decision making, Terrorism and counterterrorism, Two-stage game",
author = "Abdolmajid Yolmeh and Melike Baykal-G{\"u}rsoy",
year = "2019",
month = "3",
doi = "10.1287/deca.2018.0377",
language = "English (US)",
volume = "16",
pages = "46--66",
journal = "Decision Analysis",
issn = "1545-8490",
publisher = "INFORMS Inst.for Operations Res.and the Management Sciences",
number = "1",

}

Two-stage invest–defend game : Balancing strategic and operational decisions. / Yolmeh, Abdolmajid; Baykal-Gürsoy, Melike.

In: Decision Analysis, Vol. 16, No. 1, 03.2019, p. 46-66.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Two-stage invest–defend game

T2 - Balancing strategic and operational decisions

AU - Yolmeh, Abdolmajid

AU - Baykal-Gürsoy, Melike

PY - 2019/3

Y1 - 2019/3

N2 - Protecting infrastructures and their users against terrorist attacks involves making both strategic and operational decisions in an organization’s hierarchy. Although usually analyzed separately, these decisions influence each other. To study the combined effect of strategic and operational decisions, we present a game-theoretic, two-stage model between a defender and an attacker involving multiple target sites. In the first stage, the defender (attacker) makes a strategic decision of allocating investment resources to target sites to improve the defense (attack) capabilities. We consider two cases for investments in the first stage: (1) unconstrained and (2) budget-constrained models. The investment allocations for each target site determine its detection probability. In the second stage, the players make operational decisions of which target site to defend or to attack. We distinguish between two games that arise in the second stage: the maximal damage game and the infiltration/harassment game. We prove that the solution to each game under budget constraints is unique. In fact, when the second-stage game is of the infiltration/harassment type, the invest–defend game has a unique closed-form solution that is very intuitive. The results reveal that an increase in defense investments at a target site decreases the probability of both defending and attacking that target. However, an increase in attack investments increases the probability of both defending and attacking that target. Similarly, an increase in the defender’s (attacker’s) investment efficiency leads to a decrease (increase) in investments of both the defender and the attacker. Finally, the model is applied to real data to obtain the equilibrium investment and defense strategies. The results from real data demonstrate that the attacker’s penalty from a failed attack is an important factor in determining the defender’s optimal distribution of investments and defense probabilities. The defender’s second-stage defense decisions complement the first-stage investment decisions; that is, among target sites that receive little or zero investment, the most important one is covered with a relatively high defense probability in the second stage. Moreover, as the attacker’s budget increases, the defense investments shift from less important sites to more important ones.

AB - Protecting infrastructures and their users against terrorist attacks involves making both strategic and operational decisions in an organization’s hierarchy. Although usually analyzed separately, these decisions influence each other. To study the combined effect of strategic and operational decisions, we present a game-theoretic, two-stage model between a defender and an attacker involving multiple target sites. In the first stage, the defender (attacker) makes a strategic decision of allocating investment resources to target sites to improve the defense (attack) capabilities. We consider two cases for investments in the first stage: (1) unconstrained and (2) budget-constrained models. The investment allocations for each target site determine its detection probability. In the second stage, the players make operational decisions of which target site to defend or to attack. We distinguish between two games that arise in the second stage: the maximal damage game and the infiltration/harassment game. We prove that the solution to each game under budget constraints is unique. In fact, when the second-stage game is of the infiltration/harassment type, the invest–defend game has a unique closed-form solution that is very intuitive. The results reveal that an increase in defense investments at a target site decreases the probability of both defending and attacking that target. However, an increase in attack investments increases the probability of both defending and attacking that target. Similarly, an increase in the defender’s (attacker’s) investment efficiency leads to a decrease (increase) in investments of both the defender and the attacker. Finally, the model is applied to real data to obtain the equilibrium investment and defense strategies. The results from real data demonstrate that the attacker’s penalty from a failed attack is an important factor in determining the defender’s optimal distribution of investments and defense probabilities. The defender’s second-stage defense decisions complement the first-stage investment decisions; that is, among target sites that receive little or zero investment, the most important one is covered with a relatively high defense probability in the second stage. Moreover, as the attacker’s budget increases, the defense investments shift from less important sites to more important ones.

KW - Budget constraint

KW - Infrastructure security games

KW - Sequential game

KW - Simultaneous game

KW - Strategic decision making

KW - Terrorism and counterterrorism

KW - Two-stage game

UR - http://www.scopus.com/inward/record.url?scp=85064946189&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064946189&partnerID=8YFLogxK

U2 - 10.1287/deca.2018.0377

DO - 10.1287/deca.2018.0377

M3 - Article

AN - SCOPUS:85064946189

VL - 16

SP - 46

EP - 66

JO - Decision Analysis

JF - Decision Analysis

SN - 1545-8490

IS - 1

ER -