Research On Reduction And Core Problem Of Cooperative Network

Many systems in the real world exist in the form of networks,such as the social system of interpersonal network and the ecological system of neural network,etc.These networks all belong to the category of complex network research.Complex networks have always been the hot field of research in multi-disciplinary intersection.The re-search object of this paper is the cooperative network,which belongs to the complex networks.It is the network topology formed by the agent to the communication relation-ship.The agent with communication relationship in the cooperative network can freely cooperate(alliance),freely choose to cooperate with other agents to accomplish indi-vidual or collective goals,which overlaps with the research field of cooperative games,so this paper combines the complex network system with the research of cooperative game to solve the core of the cooperative game on the cooperative network.Recently,AI(Artificial Intelligence)and MAS(Multi-agent-system)is studying whether forming a major league(including all agent coalition),because dividing the major league into multiple small leagues may have better performance,but from the perspective of reality game,this has also caused research hotspot of multi-agent-system in recent years the problem of coalition generation under constraint conditions.The core of its research is to divide the agent set(major coalition)into several sets(coalition)of good properties,so that the sum of the benefits generated by the divided small sets is greater than the benefits of the pre-partition major league.This problem is computationally challenging.This paper mainly studies how to find the optimal coalition structure on a network of co-operation,and proposes a CNC algorithm to find the optimal joint supervision structure on the cooperation network by improving the DP algorithm.Different from the tradi-tional research on the optimal coalition structure generation problem,this paper assigns the coalition income to the agent participating in the game,and uses the definition of the kernel to describe the stable state of the cooperative game.That is,as the game time in-creases,the agent in the game gradually tends to be stable,and it is no longer frequent to convert coalition.The coalitional structure and its interest division are the stable state of game,and calculating the stability of the coalition in advance is convenient for managers to formulate planning and management,so as to better achieve their own participation in the game.The purpose is to avoid the meaningless consumption of resources.When assigning the coalition income to the agent participating in the game,the distribution according to work is adopted as the initial allocation plan,and the theory of negotiation set and stable cost in the cooperative game is taken as the initial allocation adjustment.The scheme ensures that the obtained benefit division satisfies the expected benefit of each agent in the coalition state,thereby ensuring that the optimal coalition structure generated by the CNC algorithm and its allocation stable.With the complexity of the relationship between agents in the cooperative network,the cooperative efficiency be-comes a difficult problem to extract effective information from large-scale cooperative networks.In order to obtain information more efficiently in the female complex co-operative network,an algorithm for effectively reducing complex network is proposed.The algorithm is a meaningful research work.This paper briefly discusses the feasi-bility of reducing the complexity of a large-scale cooperative networks,and designs a CNR algorithm for fast reduction of complex cooperative networks.The location of the agent in the cooperative network and their potential in the coalition is the two impor-tant prisoners that affect the status of the agent in the cooperative network.The CNR algorithm adopts the concept of location advantage and influence degree,and finds the agent with both location advantage and influence as the core agent,and starts from the core agent.The agents that facilitates the development of the coalition join,and realizes goal of reducing the complex cooperative network into multiple simple topologies,and reduces the time for obtaining effective information from the cooperative network.Finally,the analysis of these two algorithms proves that the time complexity of C-NC algorithm and CNR algorithm are O(n×3~n)and O(n~2)respectively.The feasibility of the two algorithms is verified in the experimental part.The experimental results are analyzed,and the influence of the number of participating players and the cooperative network topology on the solution time of the CNC algorithm is obtained.The number of core nodes and the number of nodes in the cooperative network are set to reduce the time required for the CNR algorithm to cooperate with the network.In the comparison experiment,the CNC algorithm is used to solve the kernel of the cooperative network before and after the CNR algorithm reduction.The comparison of the solution time and the solution accuracy verifies that the CNR algorithm can improve the speed of extracting effective information in the cooperative network.