Solving fuzzy p-hub center problem by genetic algorithm incorporating local search

The p-hub center problem has extensive applications in various real-world fields such as transportation and telecommunication systems. This paper presents a new risk aversion p-hub center problem with fuzzy travel times, in which value-at-risk (VaR) criterion is adopted in the formulation of objection function. For trapezoidal and normal fuzzy travel times, we first turn the original VaR p-hub center problem into its equivalent parametric mixed-integer programming problem, then develop a hybrid algorithm by incorporating genetic algorithm and local search (GALS) to solve the parametric mixed-integer programming problem. In our designed GALS, the GA is used to perform global search, while LS strategy is applied to each generated individual (or chromosome) of the population. Finally, we conduct two sets of numerical experiments and discuss the experimental results obtained by general-purpose LINGO solver, standard GA and GALS. The computational results show that the GALS achieves the better performance than LINGO solver and standard GA.

Figure optionsDownload as PowerPoint slideHighlights
► A new fuzzy p-hub center problem with VaR criterion in the objective is presented in the paper.
► For common fuzzy travel times, the proposed problem is turned into its parametric mixed-integer programming model.
► A hybrid algorithm by integrating GA and LS is designed to solve the equivalent mixed-integer programming problem.
► Numerical experiments demonstrated the developed modeling ideas and the effectiveness of the proposed method.