Parallelism Analysis Of Differential Evolution Algorithm Based On Graph Theory

In recent years,computational intelligence has played an important role in the development of people's life.It made a great contribution to the industrial and the scientific.In order to solve the optimization problems that generated in the industrial.Researcher mimiched from the nature to provided biology-inspired methods.Although,they have already studied the rules of these algorithms,theoretical analysis on these algorithms are few.It is important to understand in depth that theoretical analysis of algorithms can get the inherent rules.Because the simple and powerful feature,the differential evolution(DE)algorithm become an important part of random heuristic algorithm.According to the great performance of DE,we took the advantage of the principle by parallelism characteristics based on the iteration of DE algorithm.We use the knowledge of graph theory to find the reason why the random heuristic algorithm could be more stability and robustness.Therefore,it is necessary to find the support of theory analysis for random heuristic algorithm.By the knowledge of graph theory,it will directly exhibit the essential reason in the DE algorithm.To solve the above problem,the research results of this paper are three aspects:(1)From an angle of parallelism chararcteristic,the population behaviors is tend to be more mass and parallelism.Investigated the parallelism foundation in the population of algorithm iteration process.We wonder it could improve the theroy development of the DE algorithm.(2)This paper will research the parallelism characteristics based on graph theory in differential evolution algorithm.Graph theory as an important branch of mathematical science,will effective display the individual relationship in the process of the algorithm.The relationship between individual will clearly show that the parallel characteristics of the population.(3)In order to explore the reason why DE could have stable and robust quality,we analyzed the parallelism of the evolutionary process in the iteration of DE algorithm.In the evolutionary process of DE algorithm,the vertexes presented by individual in the population and the edges expressed by the relation between two individuals in the evolutionary process.This article enrich the theory of random heuristic algorithm and get the positive correlation between parallelism characteristic and the performance of differential evolution algorithms.