Clustering model and algorithm for production inventory and distribution problem

One of the most challenging problem for optimizing supply chain is the coordination of production and distribution decision. This paper considers a customer clustering problem for a large scale production, inventory and distribution problem (PIDRP) where the demands at customers need to be satisfied in each period in the planning horizon with limited production and transportation capacity. The clustering aspect of the distribution problem is an important component for reducing complexity of the dispatch operation especially when the number of customers is large.In this paper an enhanced clustering model that considers factors (e.g. demand pattern and holding costs) that could affect the operating cost throughout the planning horizon is introduced. The algorithm based on a reactive tabu search for solving the clustering problem for PIDRP is proposed. A novel feature of the algorithm is to create adaptive core clusters which are used in the clustering process instead of the original data points. With this approach the complexity of the original problem is reduced and the proposed algorithm is able to solve the PIDRP much more efficient. Computational testing on instances up to 200 customers and 20 time periods demonstrates the effectiveness of the proposed model and algorithm in term of solution quality and runtime. Special cases of the problem are also considered to provide useful insights on how to apply the model with different settings of model’s parameters.