An approach to integrate production-transportation planning and scheduling in an aluminium supply chain network

Abstract

This article is based on a real-life problem of a global aluminium supply chain network driven by an aluminium smelter. At each echelon of the aluminium supply chain network, several members are involved which are scattered around the world. Producing aluminium begins with bauxite mining. Next, aluminium oxide is made from bauxite and finally aluminium is produced from aluminium oxide. A novel type of mixed-integer decision-making model, including a time-continuous representation of the planning period, is presented. The model enables coordination of production quantities and times of all supply chain members in order to minimise production and transportation costs of the whole supply chain minus bonus payments for early deliveries which are stipulated between the supply chain network and its customers. Material flows can take place with or without temporary storage of intermediate products at supplying and/or receiving sites. Furthermore, relax-and-fix heuristics are presented. A number of randomly generated scenarios are presented to demonstrate that the heuristics can find nearly optimal solutions along with drastically reduced computation times. The relax-and-fix heuristic enables iterative planning between centralised and decentralised decision makers.

Keywords:

• supply chain network
• distribution
• scheduling
• mixed-integer programming
• relax-and-fix heuristic

Notes

Note

1. Dead weight tons (dwt) is the description of the loading capacity of cargo ships, whereas freights are measured in metric tons. One metric ton equals to one dwt (Bilgen and Ozkarahan 2007, p. 557; Stopford 2009, p. 424).