Developing scenario-based robust optimisation approaches for the reverse logistics network design problem under uncertain environments

Abstract

In the last decade, economic benefits and environmental legislation have imposed reverse logistics activities, induced by various forms of return, to organisations. Reverse logistics network design is a major strategic issue. This paper discusses scenario–based stochastic programming method and robust optimisation approaches including minimisation conditional value–at–risk (CVaR), p–robust regret and min–max regret models to find the most appropriate method dealing with uncertain environment in designing reverse logistics network. Firstly, the reverse logistics network design model is developed by using two–stage stochastic programming approach integrating CVaR in its objective function as a robustness criterion ensuring that the amount of objective function is not worse than the CVaR value with specified probability (confidence level) under all realisations. Also, the advantages of stochastic programming method are investigated respect to deterministic model under all defined scenarios. Then, the scenario–based robust optimisation methods are compared with the stochastic and deterministic ones to disclose their advantages and disadvantages.