Research And Application Of Multi-objective Optimization Method Of Discrete Structure Based On Differential Evolution Algorithm

Structural optimization has always been a major issue in scientific research and engineering applications.However,with the deepening of research and application,optimization problems have become increasingly complicated.Many problems include discrete variables and multiple optimization goals,and various engineering constraints need to be considered.Traditional structural optimization design methods are not competent,so people turn to use Self-learning and versatile intelligent algorithms solve these problems.Differential Evolution Algorithm(DE)is a general and efficient heuristic intelligent algorithm.Its direct search and highly parallel nature make it have the potential to deal with discrete variable multi-objective optimization problems.However,the standard DE algorithm still has shortcomings such as strong dependence on control parameters,slow convergence speed,and easy prematurity,and it cannot directly deal with discrete variables and multiple optimization goals.Therefore,it is necessary to improve it.The main research contents of this article are as follows:1.To improve the defects of the standard DE algorithm,the chaos initialization strategy is introduced,and two adaptive evolution strategies are proposed.Integrating the above improvements,the ACDE(Adaptive Chaotic Differential Evolution)algorithm is proposed,and compared with six typical algorithms through numerical experiments.The results show that the algorithm has strong optimization ability,fast convergence speed,good robustness,and reduces the parameters that are sensitive to the problem,which further improves the versatility of the algorithm.2.To improve the DE algorithm for discrete optimization,An improved discretization method and mixed population strategy are proposed.Integrating the above improvements,the ACDDE(Adaptive Chaotic Discrete Differential Evolution)algorithm is proposed and compared with eight existing algorithms through three discrete truss optimization problems.The results show that,while ensuring the optimal solution quality,the ACDDE algorithm is significantly more efficient than other algorithms.3.To improve the DE algorithm for multi-objective optimization,an improved constraint processing method is introduced,an adapted adaptive strategy and an improved diversity maintenance method are proposed.Integrating the above improvements,the IMO-ACDDE(Improved Muti-Objective Adaptive Chaotic Discrete Differential Evolution)algorithm is proposed and compared with existing algorithms through two multi-objective discrete structure optimization problems.The results show that the IMO-ACDDE algorithm can effectively deal with multi-objective layout optimization problems,and can maintain the convergence,diversity and uniformity of the optimal solution set in complex problems.4.Taking the roof frame of a passenger car as the optimization object,and taking the structural weight,torsion resistance and bending resistance as the optimization objectives,the IMO-ACDDE algorithm was used to optimize the multi-objective layout problem.Three representative optimal solutions are selected for analysis.The results show that each optimal structure is clearly laid out and arranged reasonably.While there is a large performance improvement,fewer types of cross-sections are selected,and engineering constraints such as displacement,stress,symmetry,consistency and installation location are satisfied.The results can be used directly in manufacturing without manual modification.In summary,this paper studies and applies the multi-objective optimization method of discrete structures based on differential evolution algorithm.The IMO-ACDDE algorithm proposed in this paper can simultaneously handle continuous and discrete variables,realize the overall optimization of the layout optimization of structural size,shape and topological variables,can effectively deal with multi-objective problems,and can flexibly set constraints according to engineering requirements,thus can be widely used in engineering structure optimization problems.