A hybrid BPSO approach for fuzzy facility location problems with VaR

In this paper, a fuzzy facility location model with Value at Risk (VaR) is proposed, which is a two-stage fuzzy zero-one integer programming. Since the fuzzy parameters of the location problem are continuous fuzzy variables with an infinite support, the computation of VaR is inherently an infinite-dimensional optimization problem, which can not be solved analytically. In order to solve the model, first of all, the objective function VaR is approximated through discretization method of fuzzy variables. Therefore, the original problem is converted to the task of a finite-dimensional optimization. Then, a hybrid heuristic algorithm integrating binary particle swarm optimization (BPSO), simplex algorithm and the approximation approach is designed to solve the location model. Finally, a numerical example is provided.