Towards Diverse and Affordable Route Portfolios for Industrial VRPs
This work proposes a decision-support approach for the Periodic Capacitated Vehicle Routing Problem with Time Windows and Driver Consistency, generating a small set of feasible, near-optimal solutions in cost, yet diverse in overtime distribution across drivers.
Algorithmically, the proposal integrates into an existing industrial process built on a two-phase metaheuristic (parallel GRASP construction and VND improvement) and incorporates a diversity-aware acceptance rule, requiring each solution to satisfy both a near-optimality threshold and a minimum separation from previously accepted alternatives.
The result is a portfolio of cost-comparable plans with different overtime allocations, enabling planners to select the option that best matches potentially uncertain or changing driver preferences.
Keywords: Uncertainty VRP Diversity Route Planning