Research On Optimal Design Of Robust Network Structure For Network Games

Many interacting systems in nature and human society can be modeled as networks with complex structures.Therefore,it is of great practical importance to study the nature of these networks,where network robustness depicts the performance of network systems from a macroscopic perspective,focusing on the integrity of the structure in the case of network attack and defense;the cooperation ratio of nodes analyzes the network from a microscopic perspective,focusing on the social behavior of members in the case of network games.In recent years,the optimal design of robust network structures has seen new developments due to the enhancement of optimization algorithms,which are capable of designing robust networks in certain special cases.However,there are still many shortcomings in the existing research: first,the network attack method is relatively simple and cannot effectively model the attacker’s attack strategy in the actual network attack and defense confrontation;second,the network structure design is only considered in the macroscopic perspective,but not effectively combined with the microscopic perspective of social behavior analysis in the network game.To address the above shortcomings,the main work and contributions of this paper are as follows:(1)To address the shortcomings of existing network structure optimization design work in which malicious attack strategies are limited to traditional degree and memo number attack methods,this paper proposes two attack methods based on mixed integer programming and forbidden search for solving the attacker’s optimal attack strategy on the network.To defend against the attacks,the defender uses an evolutionary algorithm for the optimal design of the network structure,while the performance of the algorithm is compared and analyzed with other traditional algorithms using network robustness evaluation metrics.The experimental results show that the robustness of the network structure optimally designed by the evolutionary algorithm is significantly better than other traditional algorithms.In general,this paper improves the shortcomings of working the attacker side of robust network structure optimization design under offensive and defensive adversarial games.(2)To address the situation that the existing network structure optimization design work does not consider the network robustness and node cooperation behavior in the evolutionary game,this paper uses a multi-objective evolutionary algorithm for network optimization design,aiming to design a network with both robustness and high node cooperation ratio.In the analysis of node cooperative behavior in evolutionary games,the correlation between network structure parameters and cooperative behavior in evolutionary games is investigated,and it is found that network heterogeneity has a positive correlation with cooperative behavior.When considering network robustness in offensive and defensive confrontations,the effect of network structure damage on the cooperation ratio of member nodes is investigated,and it is found that more efficient malicious attack methods reduce the cooperation ratio of nodes in the network.Experimental results show that the performance of optimally designed network structures on synthetic and real networks is improved in both macroscopic and microscopic aspects.Overall,this paper extends the research on robustness under evolutionary games to the optimal design of robust network structures under network evolutionary games and proposes a method to optimally design a robust network with a high node cooperation ratio.