基于PSO神经网络的进化博弈研究

针对进化博弈中博弈人是有限理性的,提出了一种基于粒子群神经网络的进化博弈决策机制。该机制将神经网络技术引入到进化博弈中,并采用粒子群优化算法(PSO算法)来训练神经网络,因而可利用神经网络来模拟博弈人在进化过程中的学习和策略调整。利用该机制分别对有限理性条件下的鹰-鸽博弈和重复囚徒困境博弈进行了研究。实验表明:PSO神经网络可以准确地模拟进化博弈中博弈人的动态学习与决策过程,能有效地指导博弈人的策略选取,是进化博弈分析的有力工具。

Study on evolutionary games based on PSO-neural networks

Considering the bounded rationality of players in evolutionary games,a decision method based on the particle swarm optimization neural network for evolutionary games is presented.The neural network is introduced into evolutionary games and is trained by the particle swarm optimization algorithm(PSO),thus this method can use neural networks to simulate the learning and strategy-choosing of the player.The hawk-dove game and the iterated prisoner's dilemma are analyzed respectively.Experimental results show that the PSO-neural network can accurately simulate player's dynamic learning and guide players to choose the best strategies,and it provides a good means for evolutionary games.