FEMA-CL: Fair Efficient Multi-Agent Course learning

Abstract Sociology shows that blindly pursuing the fairness of resource distribution will significantly reduce people’s enthusiasm for work, which is not conducive to increase total social material resources. Promoting system in stages, is, achieving on premise a certain basis, can only ensure efficiency accumulation, but also promote system. Therefore, inspired by above, we introduced multi-stage curriculum learning into fair policy multi-agent systems, and proposed novel Fair Effective Multi-Agent Curriculum Learning (FEMA-CL). The course progressively promotes large-scale systems through three stages: selfish stage, soft stage global stage. Our method easy learn efficiency, has carried out extensive experiments typical scenarios. Compared with current popular our superior performance.