Solving a Vehicle Routing Problem with uncertain demands and adaptive credibility thresholds

Abstract

Vehicle Routing Problem is an optimization problem of great interest in many real world scenarios such as waste management, delivery routing, etc. When employed in application contexts, there are several constraints that have to be considered such as vehicle fleet capacity and time windows. In real-world cases, it is common to find uncertainty in some parameters such as customer demands or route costs. This work proposes a way to address demand uncertainty based on fuzzy logic and adaptive credibility thresholds joined to a memetic algorithm to find the route assignments with minimum total cost. This approach is tested over several fuzzified benchmark instances and case scenarios to validate its adequacy and performance.