Α-Satisfactory Goal Programming Method for Multi-Objective Optimization with Priorities and Fuzzy Parameters.
Chaofang Hu,Na Yang,Na Dong,Yimin Zhou,Baocang Ding
DOI: https://doi.org/10.3233/jifs-18904
2019-01-01
Journal of Intelligent & Fuzzy Systems
Abstract:In this paper, the a-satisfactory goal programming (GP) method is proposed for multi-objective optimization (MOO) with priorities and fuzzy parameters. Fuzzy parameters are treated as fuzzy numbers, and all objectives are modeled into fuzzy goals over alpha-level sets. The order of alpha-satisfactory degrees is applied to preemptive priority requirement, where the objectives with higher priority can achieve the higher alpha-satisfactory degree. In order to guarantee the feasibility and seek the preferred solution, a priority variable is introduced to relax the strict order. For three fuzzy relations, GP is combined with the relaxed order constraint to formulate the different alpha-satisfactory optimization models. By regulating optimization parameters lambda and alpha, the most satisfactory solution over fuzzy parameters can be obtained, and the balance between optimization and priority can be realized. The reformulated alpha-GP models are proved to be feasible, and their solutions are guaranteed to be M-alpha-Pareto optimal by the test model. In order to decrease optimization burden, the algorithm to compute the alpha-maximum regulating parameter is proposed, by which the bound of lambda can be determined. The effectiveness of the proposed method is well demonstrated by numerical examples.