A hybrid algorithm for stochastic multiobjective programming problem

Abstract

The traditional approach in the solution of stochastic multiobjective programming problem involves transforming the original problem into a deterministic multiobjective programming problem. However, due to the complexity in practical application problems, the closed form of stochastic multiobjective programming problem is usually hard to obtain, and yet, there is surprisingly little literature that addresses this problem. The principal purpose of this paper is to propose a new hybrid algorithm to solve stochastic multiobjective programming problem efficiently, which is integrated with Latin Hypercube Sampling, Monte Carlo simulation, Support Vector Regression and Artificial Bee Colony algorithm. Several numerical examples are presented to illustrate the validity and performance of the hybrid algorithm. The results suggest that the proposed algorithm is very suitable for solving stochastic multiobjective programming problem.