Improving Adaptive Differential Evolution And Its Application

Differential Evolution(DE)is a swarm intelligence optimization algorithm,which has been widely used to solve practical optimization problems in recent years.DE algorithm has many advantages,such as fewer parameters and better optimization ability.However,there are still some shortcomings in dealing with some complex optimization problems,such as slow convergence speed and easy to fall into local optimum,which lead to poor solution effect.In this paper,DE algorithm is improved from the aspects of accelerating the convergence speed and reducing the probability of falling into the local optimum,and the improved algorithms are applied to solve the problem of chemical parameter identification.The specific work is as follows:(1)The fundamental idea,main operations of DE algorithm and other two kinds of swarm intelligence optimization algorithms were summarized.Aiming at the shortcomings of traditional DE algorithm,the research status and improvement direction of DE algorithm are analyzed.(2)Aiming at the slow convergence rate of traditional DE algorithm in solving complex optimization problems,this paper presents An adaptive bare-bones differential evolution based on cosine(CABDE)was proposed,which a new adaptive mechanism for mutation strategy selection was designed.The mechanism combined the Gaussian mutation strategy and the DE/current-tobest/1 mutation strategy,and employed a cosine adaptive factor to select the proper mutation strategies.The cosine adaptive factor was adjusted according to the increase of the iterations.Therefore,CABDE can adaptively choose the suitable mutation strategies in the different evolutionary stages,which can accelerate the convergence speed as well as maintain the population diversity.The performance of the improved algorithm was tested on 18 test functions.The experimental results showed that the overall CABDE algorithm accelerates the convergence speed and improves the optimization performance.(3)Aiming at the problem that traditional DE algorithm is easy to fall into local optimum when solving complex optimization problems,which results in inaccurate solution accuracy.A differential evolution with combined triangular mutation strategy(CTMDE)was proposed.CTMDE incorporated the combined triangular mutation strategy and the DE/current-to-pbest/1 mutation strategy to improve the performance.In the proposed CTMDE,the combined triangular mutation strategy introduces the combination weight to adaptively employ the information of the better individual,the general individual,and the current individual to maintain the population diversity.Meanwhile,the DE/current-to-pbest/1 mutation strategy has good local search ability and can accelerate the convergence rate,which can also decrease the probability of falling into local optimum to some extent.Therefore,The combination of the two mutation strategies forms complementary advantages in search ability,and CTMDE can keep a balance between exploration and exploitation in the search process.The performance of the algorithm was tested by 18 test functions,and the results showed that CTMDE has better optimization performance.(4)In order to further test the performance of CABDE and CTMDE in this paper,CABDE and CTMDE were used to identify the parameters of the methanol-to-hydrocarbons process,and compared with other algorithms.The experimental results showed that CABDE algorithm and CTMDE algorithm are more competitive in solving this problem.Moreover,CTMDE algorithm is better than CABDE algorithm.