A hybrid heuristic approach for production planning in supply chain networks

Abstract

Planning distributed manufacturing facilities is one of the most challenging tasks in the supply chain management. This paper proposes a production planning algorithm for the multi-level, multi-item capacitated lot-sizing problem (MLCLSP) in a supply chain network that takes back order into account. MLCLSP is a mixed integer linear programming (MIP) problem and is NP-hard. This paper presents an efficient, hybrid, heuristic algorithm named greedy rolling horizon search (GRHS) that combines a rolling horizon local search heuristic with an exact linear program (LP) solver. Computational experiments show that GRHS performs well in terms of total costs and computational time and is superior to existing meta-heuristics, such as tabu search, simulated annealing, and genetic algorithms.