Modified JSO Algorithm And Its Research And Application In Constrained Optimization Problems

Optimization problems are common problems in the fields of mathematics,engineering technology,operations research,and computer science.Evolutionary algorithms are widely used to solve optimization problems because they do not require conditions such as continuous,differentiable,and derivable,and it is able to maintain the diversity of the population so that it is not easy to get in the local optimal solution.Differential evolution algorithm is a simple and effective random algorithm based on population.It uses differential mutation operator and crossover operator to generate new offspring,and generates new individuals by the way of survival of the fittest.In view of the simplicity and efficiency of the algorithm,and its excellent performance in previous evolutionary algorithm competitions,it has received more and more attention and research from researchers.Therefore,this is very important for the improvement of DE algorithms and its application to complex problems.The article addresses the problem of premature integration of jSO algorithms in search fields of different sizes and falling into local perspectives in the evolution process,and the problem of population reduction.From these three aspects,the jSO algorithm is improved,and an improved jSO algorithm "MjSO" is proposed.Then the application of MjSO algorithm in constrained optimization is further researched.The main work of the paper is as follows: First,the MjSO algorithm is proposed.The algorithm first introduces the parameter control strategy of symmetric search to balance the exploration and development in the process of algorithm evolution.Secondly,a flexible strategy of parameters based on cosine similarity for the weight coefficient is adopted to maintain the algorithm's search capability.Finally,a new restart mechanism based on the opposite study is introduced to prevent the algorithm from falling to local acceptability.Second,the effectiveness of these improvements is to evaluate the performance of the algorithm through population cluster analysis and population diversity measurement.Third,compare the three most advanced DE variant algorithms with two original algorithms.CEC2017 Benchmark test functions and four classic engineering design problems were compared and simulated experiments.Finally,through the analysis of the results produced by the above-mentioned comparative simulation experiments,it is found that the MjSO algorithm has excellent performance compared with other comparative algorithms in terms of solution quality and convergence speed,which fully proves the effectiveness and applicability of the MjSO algorithm.In the future,it is very promising that it can be regarded as an effective auxiliary tool for more complex optimization models and scenarios.