Research On Engineering Design Optimization Problem Based On Pyramid Evolutionary Strategy

In the engineering design optimization problems,the optimal solution can improve the design quality of complex engineering system and reduce a lot of cost consumption.Therefore,it is of great significance to study the optimization of engineering design.At present,the swarm intelligent algorithms are used to solve the optimization problem of engineering design,engineering system design problems are mostly mixed-integer nonlinear programming problems,the traditional swarm intelligent algorithms fail to solve the contradictions between exploitation and exploration,competition and cooperation among individuals or populations when solving mixed integer programming problems,but pyramid evolution strategy has clear division of labor mechanism and promotion mechanism,which can solve the two contradictions,therefore,this paper studies PES algorithm and applies it to solve engineering design optimization problems.The main work is as follows:In order to optimize the problems more efficiently,in this paper,PES algorithm is used to solve mixed integer programming problems for the first time.The test functions of mixed integer programming were simulated,the experimental results show that PES algorithm has higher success rate and convergence accuracy,they also verify the effectiveness of PES algorithm by comparing the results of the improved particle swarm optimization algorithms(CLSPSO,CLSPSO2)and the improved differential evolution algorithms(rid DE,rid DE2).However,further research founds that PES algorithm is easy to fall into local optimum when solving engineering design optimization problems,in view of this defect,this paper proposes a pyramid evolution strategy based on unimodal normal distribution acceleration.The UNDA-PES algorithm has three changes compared with PES algorithm,first,changes in the way of the generation of initial population,in order to reduce the search time of the algorithm and guide the direction of population evolution,a certain proportion of the feasible solution in the initial population is set.Second,the change of search strategy in the layer.A search strategy based on the interaction of elite and random individuals is used to update the population.Thus,the convergence speed of the algorithm to the optimal solution is accelerated.Third,the change of acceleration strategy within the layer,in order to avoid the algorithm falling into local optimum,the acceleration operation based on unimodal normal distribution is proposed.In addition,the convergence of UNDA-PES algorithm is analyzed,which makes the improved algorithm theoretically more complete.The UNDA-PES algorithm is applied to five well-known engineering design problems,and numerical experiment is simulated on the test function set of mixed integer nonlinear programming problems,this paper makes a comparison between PES algorithm and the latest swarm intelligence algorithms.The experimental results show that the UNDA-PES algorithm has faster convergence speed and higher convergence accuracy in solving engineering design problems.