协同免疫量子粒子群算法求非合作博弈Nash均衡解

COEVOLUTIONARY IMMUNE QUANTUM PARTICAL SWARM OPTIMIZATION IN SOLVING NASH EQUILIBRIUM FOR NON-COOPERATIVE GAME

考虑n人非合作博弈Nash均衡求解问题。将混合策略意义下的Nash均衡转化为最优化问题;把免疫记忆、自我进化、信息共享机制加入量子粒子群算法,通过概率浓度选择公式来保持种群的多样性,提出协同免疫量子粒子群算法。4个经典的数值算例说明,该算法优于免疫粒子群算法,具有较强的寻优能力和收敛性能。

Considering Nash equilibrium for N-persons non-cooperative game,Nash equilibrium problems of mixed strategy was converted to optimization problems.We introduced immune memory,self-evolution,information sharing regulation into quantum particle swarm optimization,and maintained the diversity of the population by probability density selection.Based on this,we proposed a coevolutionary immune quantum particle swarm optimization (CIQPSO).The four classical numerical examples show that CIQPSO is superior to the immune particle swarm optimization algorithm,and has stronger optimization ability and convergence performance.