Solution of pursuit/evasion Differential Games using genetic algorithms

Abstract

Differential games (DG) uses differential equation(s) to model the state varying of two (or more) sides in games. Saddle point is the solution of DG, which is the emphasis and the difficult point in DG especially pursuit/evasion DG (PEDG) study. In this text, saddle point and its meaning was discussed, numerical solution methods of saddle point of PEDG were assorted, three kinds of models were studied while using genetic algorithms (GA) and genetic programming (GP) to solve PEDG. The correctness of one kind of these models is proved by comparing the numerical solution with the analytic solution of a simple problem.