Route Design for Lean Production Systems

We consider the problem of routing a fleet of vehicles to pick up components from a network of suppliers and to deliver them to a fixed depot in a lean production system. The pursuit of low work-in-process inventory and production smoothing throughout the system introduces complicating side constraints, creating an enriched case of the vehicle routing problem with time windows and split deliveries. We present a two-phase routing and scheduling approach to address this problem. The routing phase consists of a nested tabu search heuristic that iterates between determining the suppliers' visit frequencies and developing vehicle routes. Given a routing plan, the scheduling phase determines the timing of supplier visits using a binary integer program designed to promote production leveling. Through computational testing on real-world data sets, we compare our solutions to those in the literature and those used by practitioners in the industry.