Research On Network Reconstruction Algorithm Based On Evolutionary-game

In the natural and social systems, there is a wide variety of complex networks, and their research in recent years has also been interdisciplinary. Due to the universality and diversity of complex network, it draws much attention from research workers at home and abroad. The study of complex network not only has an important significance in real life, but also for understanding human nature and society’s long-term development.All complex networks are composed of a number of interacting individuals. The study on the interactive model of these individual patterns can not only reveal the topology of the network further, but also plays a greatly important role in the dynamics of the network. In a variety of complex networks, such as natural and social systems, the evolutionary game model is a common interactive model. In this paper, by the use of the prisoner’s dilemma data between the individuals which are on the underlying of network, we further reveal the dynamics of the network and a real social network is reconstructed. The specific work is as follows:1. Game theory is used to study the interaction network. Some individual members, team groups or other organizations of the game correspond to the equivalent nodes of the network, the game’s equivalent of the game between people connected edges between nodes. Each allowed to choose from a selection of actions or policies and implement them, each corresponding result achieved in the process is equivalent to the value of inter-node game proceeds obtained as attribute information node for network reconstruction, and get the network topology.2. A new method based on random game strategy and compressed sensing of network reconstruction is proposed. In the network reconstruction algorithm, the game data is used in the game prisoner, rather than weak prisoner game. Another game strategy updated the guidelines in this algorithm is randomly assigned to ensure that diversity policies. The same series of multi-generational game strategy game strategy or favor a single node don’t appear, and this greatly improves the efficiency of the reconstruction.3. A new method based on genetic algorithm and game of network reconstruction is proposed. Currently, the network reconstruction method is mainly based on the assumption of a network, and this is coupled to the shock. The continuous time data is then generated by the network to realize the reverse engineering of the network. And our approach is in the discrete-time case, the resulting interaction model prisoner. Game results by combining genetic algorithm is based on approximate revenue value of the network for network reconstruction, and has achieved good results.