Evolutionary Minority Game On Small-world Networks In The Disadvantaged Situation

In financial markets, agents involved in financial transactions in addition to face competitive pressures within the industry, but also have to accept the impact from the entire market. The performance of evolutionary minority game model(EMG) was investigated which introduced imitation mechanism on WS small-world network in the disadvantaged situation, explored the resource utilization of the system and the strategy changing of agents who among the network under the impact of the system’s systemic risk. Within a certain range of expressed environment caused by the system’s systemic risk, the model with imitation mechanism leads to a good optimizing effect of resource allocation, agents can learn and imitate with neighbors while on the network, improve the hold-time of strategies to relieve the impact of such bad conditions. However, when the environment is extremely bad, the interaction will reduce the efficiency of the system, agents are taking a more cautious strategy. An evolutionary minority game model with strategy periodically changing was proposed, during a period τ steps of the game, a criterion is set by the success rate with percentage ranks pr, agents who have a lower success rate than the criterion change their strategies. Through simulation it’s found that the model has a good effect optimal allocation of resources, system with lower pr is more efficient, and they are approximately linearly related within a certain range. The study also showed that with the increase of the strategy elimination rate, agents change their behaviors from obvious extreme way to a way of variety strategies, which have a uniform distribution of strategies. When the system is on WS small-world networks, the imitate mechanism between agents will help improve overall system balance evolution speed, excellent strategy to accelerate the transmission of information, and it shows a higher resource utilization compared to the self-centered(self-around) evolution strategy model and strategies random(rand) evolution model.