Multi-Objective Particle Swarm Optimization Based On Fuzzy Preferences And Its Application In Inventory Control

Particle Swarm Optimization (PSO) is a swarm intelligence optimization algorithm based on population iteration, each particle adjusts its own flying to the best position according to its previous best performance as well as the best previous performance of the whole population.With advantages of few parameters, high convergence speed, PSO are widely used in Multi-objective Optimization Problems.Integrating the preference informance can improve the performance and reduce the complexity of the algorithm,and obtain quantities of candidate solutions which is consistent with decision-maker'preference informance, However,An effectice method is desired to solve the issue of how to choose an expected solution consistent with the decision-maker'preferences from the Pareto optimal solution set.In this paper, we have done some research on the improvement of PSO,and the integration of fuzzy preference.then,the improved algorithm is applied to inventory control of the enterprise supply chain management to improve the inventory control level of the enterprise.The main contents are summarized as follows:(1) As for the defects of deficiency of population diversity and the slowness of late convergence, we introduce the adaptive Gaussian mutation and dynamic inertia weight to balance the diversity and convergence of PSO. The Adaptive Gaussian mutation is used to improve the population diversity of PSO and the Dynamic inertia weight PSO strategy to improve the convergence rate in the late process of the algorithm.The simulation experiment on six test functions shows that the improved PSO has a better performance.(2) As for the fuzzy preference of decision-maker, we propose a method named Based on Multi-objective Particle Swarm Optimization and Fuzzy Preferences(FPMOPSO),which first makes some improvements of Multi-objective Particle Swarm Optimization based on Bipolar Preferences(BPMOPSO), then uses the theory of fuzzy objective decision ,finally constructs a utility function based on fuzzy preferences in accordance with the weight of decision-makers and their preferences to objectives.The proposed algorithm can solve the problem of choosing an optimal solution from the Pareto solution set.(3)For the practical application of continuous review stochastic(r, Q) inventory model, we use the FPMOPSO to make up the defect of disability of integrating the DM'preference.Simulation result shows that the method can effectively integrate the DM'preference and reduce the searching space and improve the efficiency of the algorithm as well, finally, it can provide the optimal and interactive inventory control decision solution.