Research On Constrained Optimization Problem Based On Cellular Genetic Algorithm

In today's society,constrained optimization problems exist in all aspects of our lives,such as engineering design,image processing,scheduling problems and other fields.Generally,when dealing with constraint optimization problems,the most mainstream and effective method is to combine evolutionary algorithms with constraint processing techniques.However,in the processing process,such methods have low convergence accuracy,poor performance in high-dimensional functions and weak robustness.How to overcome these defects becomes the focus of our research.In previous research,we can know that the cellular genetic algorithm has the advantages of good diversity and robustness.However,the current research using the cellular genetic algorithm combined with constraint processing technology to deal with constraint optimization problems is very scarce.In order to effectively improve the defects of stochastic algorithm in constraint optimization,this paper proposes ?CGA algorithm and a DPCGA algorithm based on cellular genetic algorithm combined with?constraint processing technique and dual population storage technique.The ?CGA algorithm is proposed to deal with the constrained optimization problem.The cell genetic algorithm is combined with the introduced ?-constrained processing technology,and the proposed strategy of a preference index and a Cauchy mutation operator.The preference index is used in the early stage of truncated algebra to make the population effectively converge towards the feasible region,and the Cauchy mutation operator is used in the later stage of the truncated algebra to avoid the algorithm from falling into local optimality.After comparison with other algorithms,the algorithm has good convergence accuracy and performs better in high-dimensional test functions,but is less robust in some test functions.A DPCGA algorithm combining cell genetic algorithm and dual-population storage technology is proposed to deal with constrained optimization problems.During the initial population evolution,new updating strategies for feasible and infeasible individuals are proposed;when the feasible individuals are substituted into the cell structure and then evolved,the mutation operator based on the concept of particle swarm and the constraint handling restrictions measure are proposed to increase the population diversity.Finally,compared with other algorithms,the results show that the algorithm is not ideal in the test function with multiple local optimal and low objective function dimensions,but in other test functions,the algorithm not only has higher convergence accuracy,robustness and performance in high-dimensional functions are better than other algorithms.Aiming at the problem of constraint optimization,this paper proposes the ?CGA and DPCGA algorithms that use cellular genetic algorithm as an evolutionary algorithm combined with constraint processing technology.Finally,it is concluded that using cell genetic algorithm as evolutionary algorithm to handle constrained optimization problems has better convergence accuracy and performs better than other algorithms in high-dimensional function tests.However,the two algorithms have the disadvantages of poor robustness and poor applicability,which can be used as the focus of future research and improvement.