A Genetic Programming Approach to Strategy Optimization in the Extended Two-Dimensional Pursuer/Evader Problem

This paper describes a genetic programming system that evolves optimized solutions to the extended two-dimensional pursuer/evader problem. The pursuer/evader problem is a competitive zero-sum game in which an evader attempts to perform maneuvers to escape a faster, more agile pursuer. The extended problem is more realistic than previous formulations because the evader and pursuer are modeled as point masses that are capable of limited thrusting and turning forces, and are subject to drag forces and momentum. The pursuer initially aims at a predicted capture point, and uses proportional navigation to attempt to maintain a constant line-ofsight angle with the evader. The game ends favorably for the evader if it manages to stay outside the lethal radius of the pursuer for the duration of the encounter (limited by the effective range of the pursuer). To solve the extended two-dimensional pursuer/evader problem, a strategy must be identified by which an evader (such as an F-16C fighter aircraft) may maneuver to successfully evade pursuers (such as surface-to-air missiles) starting from a wide range of potentially lethal relative initial positions.