A new heuristic approach for a multi-depot three-level location-routing-inventory problem

The proposed approach investigates a three-level supply chain. It consists of a factory with uncapacitated inventory that produces one type of product at the first level, several distribution centres at the second level that act as distributors of this product, and a group of customers scattered in different geographical locations at the third level. While the location of the customers is predetermined, the location of the distribution centres needs to be determined. Since the mathematical model for this problem is NP-hard, a meta-heuristic method proposed to solve medium and large size problems using a three-stage algorithm. The first stage generates initial location and routing using limited local approach and modified algorithm. The second stage repeatedly improves location and routing decisions, while in the last stage, algorithms effectively search for better solutions by including diversification and intensification for solutions.