Research On Stochastic Multi-objective Simulation Optimization Method Based On Improved MOCBA Strategy

In real life,many engineering science and application problems have randomness and multi-objectives,which poses great challenges to the optimization of these problems.In recent years,people have begun to consider using simulation methods to solve stochastic multi-objective optimization problems.In the use of simulation methods to solve stochastic multi-objective optimization problems,how to improve the efficiency of simulation is a major factor to consider,so some scholars have proposed a Multi-objective Optimal Computing Budget Allocation(MOCBA).The MOCBA strategy improves the efficiency of the simulation.However,in the simulation resource allocation,only the dominance of the set is considered and the distribution is not considered.In solving the stochastic multi-objective problem,people want to obtain a more uniform and more distributed set of solutions.Therefore,based on the existing simulation optimization method,this thesis proposes a strategy to improve the existing simulation optimization method,and proposes the concept of contribution degree,and integrates the contribution factor into the existing simulation optimization method.The improved simulation optimization strategy is combined with the multi-objective optimization algorithm,and applied to the stochastic multi-objective test function and the uncertain traveling salesman problem to carry out simulation experiments.The sensitivity analysis proves that the improved algorithm has strong sensitivity to continuous random multi-objective problems.The accuracy analysis proves that the accuracy of the solution set obtained by multi-objective problems decreases with the increase of the uncertainty of the multi-objective problem;Through the performance index compared with the original algorithm and the existing algorithm,it is proved that the improved algorithm can increase the distribution of the solution set.The specific research content includes the following aspects:1)Research on improved MOCBA strategy based on contribution.By analyzing the original MOCBA strategy and aiming at its lack of distribution in the solution set,the concept of contribution is used to utilize each simulation information of the simulation scheme.Experiments show that the improved MOCBA strategy can improve the distribution of the solution set.2)Combining the improved MOCBA strategy with the evolutionary multi-objective algorithm NSGA-?,a stochastic multi-objective optimization algorithm based on improved MOCBA strategy is proposed.The effectiveness of the proposed algorithm for solving stochastic problems is verified by the proposed algorithm for stochastic multi-objective test functions.The sensitivity analysis proves that the improved algorithm is strong for continuous stochastic multi-objective problems.The Accuracy analysis proves that as the uncertainty of multi-objective problems increases,the accuracy of the solution set obtained decreases.3)The improved algorithm is applied to the uncertain multi-objective traveling salesman problem.By comparing the improved algorithm,the original algorithm and the existing algorithm,it is verified that the improved algorithm can improve the distribution of the solution set of the stochastic multi-objective problem.