ECAI-2002 Conference Paper

[full paper]

Combining Two Heuristics to Solve a Supply Chain Optimization Problem

Hoong Chuin Lau, Yuyue Song

In this paper, we consider a real-life supply chain optimization problem concerned with supplying a product from multiple warehouses to multiple geographically dispersed retailers. Each retailer faces a deterministic and period-dependent demand over some finite planning horizon. The demand of each retailer is satisfied by the supply from some predetermined warehouse through a fleet of vehicles which are only available within certain time windows at each period. Our goal is to develop a co-ordinated schedule such that the system-wide total cost over the planning horizon is minimised. This problem in essence is an amalgamation of two classical NP-hard optimizatin problems: the Dynamic Lotsizing problem and the Vehicle Routing problem. In this paper, we propose an efficient rolling forward heuristic that combines two heuristics to solve this problem. Numerical experiment results show that our approach can achieve, on average, within 10% of the lower-bound proposed by by Chan, Federgruen and Simchi-Levi (1998) for some specific instances generated from Solomon benchmarks.

Keywords: Scheduling, Planning, Meta-Heuristics for AI, Search

Citation: Hoong Chuin Lau, Yuyue Song: Combining Two Heuristics to Solve a Supply Chain Optimization Problem. In F. van Harmelen (ed.): ECAI2002, Proceedings of the 15th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2002, pp.581-585.