非完备策略的演化少数者博弈模型研究

Study of Evolutionary Minority Game Model with Incomplete Strategies

提出了一种基于非完备策略的演化少数者博弈模型,它在演化的同时允许参与者的策略包含部分随机位,并且规定当主规则随机选择时,由次规则顶替指导。对比分析和数值模拟表明,新模型由于在策略结构上采用“缺席的等级制度”,其性能相对于普通演化MG模型有显著提升,能够以更小的记忆步长和更稳定的策略组成,进化到一个近乎理想的协作状态。

This paper proposes a new model of evolutionary incomplete minority game （EIMG）, which features a default hierarchy of rules with evolution. It introduces random bits into strategies of agents and is capable of applying the secondary rule in the absence of the primary one. Analysis of the numerical experiment results indicates that, in comparison with the evolutionary minority game （EMG） model, the EIMG model can greatly improve the overall performance and evolve to an approximate ideal status very soon with less memory steps and more stable combination of strategies.