A risk-based game theory model of navy and pirate behaviors

Abstract

Pirate attacks continue to plague the maritime industry, causing incalculable damage to property and persons, and necessitating naval escorts through dangerous waters in order to protect shipping. In this paper, we present a two-stage analytical framework for analyzing the interactive relationship between pirates and naval groups. In the first stage, we construct a Bayesian network (BN) in order to calculate the rate of successful hijackings. We then conduct sensitivity analysis of the root nodes and introduce two hypotheses with which to increase the accuracy of the BN. The results show that the presence of a naval escort, and the seasonal quarter are among the primary factors that determine the rate of pirate hijackings. In the second stage, after we conduct the sensitivity analysis, we use the results of the BN as a key input into a Nash game theoretic model (in which pirates and navy groups are the players), in order to explicate the expected payoffs of each party. We then calculate the equilibrium attack rate by solving the Nash equilibrium point equations. In the final section, we draw meaningful insights from a numerical study in order to provide stakeholders with decision-making support in different situations.