Co-Evolutionary of Public Goods Games Based on Learning Topology

Network evolution is closely related to nodes revenue, which is determined by the strategies and topology of all participants in the group, topological structure and strategy interact with each other in the process of game. We want to know how the topology structure of nodes will affect the strategy evolution after adjusting it according to certain rules. According to this, we propose an edge-breaking and re-connection mechanism based on the learning node topology. In the learning process, the nodes can not only learn the strategies of the neighbor nodes, but also learn the topology structure of the neighbor nodes through the excess average degree. Each node relies on its own expectation degree to keep in line with the excess average degree of the learned node. Through Monte Carlo simulation experiments, we found that the cooperation starts when the gain factor is small, and soon the cooperation level can be high when the gain factor is increasing. This indicates this edge-breaking and re-connection mechanism based on learning other node’s topology can effectively promote the generation and maintenance of cooperation.