Research On Improved Genetic Algorithm Based On Multi-objective Optimization Problem

With the development of human society,many decision-making problems in the real world tend to become more complicated.The mathematical model constructed with a single goal is often not enough to describe all the characteristics of the problem.Therefore,the mathematical model of the actual problem often contains multiple solving goals,that is,multi-objective optimization problem,or MOPs.In order to solve this problem,scholars at home and abroad have proposed a series of advanced algorithms based on the basic genetic algorithm,including simulated annealing algorithm,fast non-dominant(NSGA-?)algorithm,target constraint algorithm,x-constraint algorithm,Apriori algorithm,gradient tree enhancement GTB algorithm and gradient descent algorithm.Among these algorithms,the optimization efficiency of NSGA-? algorithm for multi-objective problems has achieved great success.However,the local search capability of NSGA-? algorithm is affected by crossover and mutation operators.After the selection,mutation operator was improved,as well as to the tournament selection method modified ISMNSGA-? algorithm while it is possible to make up for the deficiency of the NSGA-? algorithm,but its time complexity is higher,and as a multi-objective optimization target dimension of the rights to add,the performance began to decline,moreover,the algorithm only adopts one crossover and mutation from beginning to end,which is easy to fall into local optimization.In order to solve these problems,the adaptive strategy is adopted,which can avoid the algorithm falling into local optimization.A differential evolutionary algorithm(DENSGA-?)based on L-nearest neighbor distance is proposed.Two populations were created in DENSGA-?,and the competition between the two populations was used to select the better individuals to enter the next generation population.Finally,by comparing the experimental results with the original NSGA-? and ISMNSGA-? algorithms,the experimental results prove the effectiveness of the algorithm,and the search process achieves a balance between the global search and the local search.