Optimal game theoretic solution of the pursuit‐evasion intercept problem using on‐policy reinforcement learning

Abstract This article presents a rigorous formulation for the pursuit‐evasion (PE) game when velocity constraints are imposed on agents of or players. The is formulated as an infinite‐horizon problem using non‐quadratic functional, then sufficient conditions derived to prove capture in finite‐time. A novel tracking Hamilton–Jacobi–Isaacs (HJI) equation associated with value function employed, which solved Nash equilibrium policies each agent arbitrary nonlinear dynamics. In contrast existing remedies proof PE game, proposed method does not assume players moving their maximum velocities and considers priori. Attaining optimal actions requires solution HJI equations online real‐time. We overcome this by presenting on‐policy iteration integral reinforcement learning (IRL) technique. persistence excitation IRL work satisfied inherently until occurs, at time ends. Furthermore, backstepping control track desired trajectories generalized Newtonian Simulation results provided show validity methods.