Tabu search heuristic for inventory routing problem with stochastic demand and time windows

Abstract

This study proposes the hybridization of tabu search (TS) and variable neighbourhood descent (VND) for solving the Inventory Routing Problems with Stochastic Demand and Time Windows (IRPSDTW). Vendor Managed Inventory (VMI) is among the most used approaches for managing supply chains comprising multiple stakeholders, and implementing VMI require addressing the Inventory Routing Problem (IRP). Considering practical constraints related to demand uncertainty and time constraint, the proposed model combines multi-item replenishment schedules with unknown demand to arrange delivery paths, where the actual demand amount is only known upon arrival at a customer location with a time limit. The proposed method starts from the initial solution that considers the time windows and uses the TS method to solve the problem. As an extension, the VND is conducted to jump the solution from its local optimal. The results show that the proposed method can solve the IRPSDTW, especially for uniformly distributed customer locations.

Keywords

• inventory routing problem
• stochastic demand
• time windows
• tabu search
• variable neighborhood descent