NEUTRAL AND RISK-AVERSE MEDIAN P-HUB LOCATION PROBLEM WITH MULTIPLE ALLOCATION

This article proposes a formulation for the p-hub median location problem with multiple allocations(UMApHMP) under uncertainties in costs and flows. Neutral risk (NEU) stochastic models based on expected value and risk-averse models utilizing Conditional Value at Risk (CVaR) and Mean-Conditional Value at Risk (MCVaR) are proposed, applied to the UMApHMP hub problem. The mathematical formulations are presented and computational experiments are conducted in CPLEX using instances from the literature. An analysis was carried out on the impacts of the uncertainty level, the choice of risk function, the sensitivity to the parameters of MCVaR, and a stochastic evaluation of the solutions obtained through simulations. The results show that the runtimes to solve the NEU, CVaR e MCVaR-based models are quite high when compared to the deterministic model. Furthermore, the results confirm that the optimal configurations for the hub network depend on the approach to dealing with uncertainties.