Community Detection Based On Evolutionary Algorithm In Complex Networks

In recent years, more and more researchers have been focus on complex networks, a large number of scientific workers work in the field of complex networks. Community structure is one of many important properties of complex networks, it reveals the behavioral characteristics and the hidden rules of the complex networks. Many systems in the real world, such as cell phone networks, the traffic road networks and so on, can be represented as the complex networks. In the network, the nodes represent objects in the systems, the edges of the network reprsent the relationships among these objects in the systems. Therefore, the study of community detection problem can help us to analyze the relationships between the various objects in the networks. Complex networks community detection problem as an important research direction is very important and very meaningful thing.Genetic algorithms and ant colony algorithms are hot research topic in the field of evolutionary algorithms in recent years, both algorithms have global search ability of populations and local search ability of the individual. They can solve some problem areas and have more advantages than the traditional optimization algorithms. Genetic algorithm has been used to complex networks community detection problem, and it achieved good results. Ant colony algorithm is seldom used in complex networks community detection problem, the use of ant colony algorithm to detect the community structure of complex networks is very meaningful. In this paper, a complex networks community detection algorithm based on improved genetic algorithm has been proposed.This paper puts forward a complex networks community detection method based on decomposition and multi-objective ant colony optimization algorithm. The mainly works in this paper are introduced in detail as follows:1. A complex networks community detection method based on an improved genetic algorithm has been proposed. Firstly, a framework of genetic algorithm is adopted,taking modularity as the objective function to detect the community structure of the network. Secondly, an improved method of initialization, a new effective local search method and the elite preservation strategy are introduced. Through the simulation experiments and compared with other algorithms, the proposed algorithm has goodresults in detecting the community structure of complex networks.2. A complex networks community detection method based on decomposition and multi-objective ant colony optimization algorithm has been proposed. Firstly, a new framework of multi-objective ant colony algorithm suitable for complex networks clustering is developed, in which, two objective optimization problems can be decomposed into a series of subproblems and each ant is responsible for one single objective subproblem and it targets a particular point in the Pareto front. Secondly, a problem-specific individual encoding method based on graph is introduced. Moreover, a new efficient local search algorithm is designed in order to improve the stability of the algorithm. The proposed algorithm has been tested on both benchmark networks and four real-world networks, and compared with other four algorithms, experimental results show that our algorithm can obtain a competitive performance for the community detection problems.