Research Of Differential Evolution Algorithm And Social-spider Optimization Algorithm

Differential evolution algorithm has become a classical method to solve the continuous numerical optimization problem.In the first part of this article,we propose a variant of DE(Differential Evolution)algorithm that we incorporate DE with the crossover thought of SSO(Simplified Swarm Optimization)algorithm and covariance matrix learning,which is called SCDE algorithm.As we know,mutation operation plays a very important role on the performance of DE.However,traditional mutation strategies of DE are all use relative position to generate a candidate vector.We try to utilize the absolution position of individual historical optimal solution to induce variation.This will greatly enhance the ability to jump out of local populations.Furthermore,the crossover and mutation operator are executed in an appropriate coordinate system which is generated by the covariance matrix of the population and it can make the crossover and mutation operation rotationally invariant.In addition,The experimental results show that the proposed operator significantly improves DE performance on a set of 28 test functions in CEC 2013 benchmark sets.And the SCDE algorithm is success to solve the TSP problem which is one of the combinatorial optimization problem.The social-spider optimization(SSO)algorithm is a new optimization technique based on the cooperative behavior of social-spiders first proposed by Cuevas.Numerical simulation results have shown that,compared with PSO,ABC,it is better powerful ability of global optimization.However,there are still some deficiencies in search strategy of SSO algorithm which has led to imbalances between exploration and exploitation which are important for optimization algorithms based on swarm intelligence.Inspired by PSO and DE,an improved SSO algorithm is proposed(denoted as wDESSO)in the second part of this paper.The main research work includes:1.a weight factor changing with iteration is introduced to control and adjust the search scope of the social-spiders;2.after social-spiders have completed their research,a mutation operator is adopted for strengthening the ability of global search and jumping out of local optimization.According to the different mutation strategies,the improved algorithms can be further classified into wDES SO-? and wDESSO-? algorithms.A group of experiments are used to verify the proposed algorithms using a set of standard benchmark functions.One compares with some competitive optimization algorithms,including original SSO,ABC and PSO,and the Wilkerson signed rank test was performed on the results of result experiment 1,the other compete with some improved optimization algorithms.The experimental results suggest that better performance of wDESSO algorithms has been achieved than other optimization algorithms based on swarm intelligence in solving complex numerical problems.