Modeling and Solving the Practical Production Planning Problem

This paper presents a mathematical programming formulation for lot sizing that is inspired from tire production practice. A mixed-integer programming based fix-and-relax heuristic approach is developed to address this problem. The time-oriented decomposition in this approach is designed to be variable dependently to obtain high quality solutions of the submodel. A numerical study using randomly generated benchmark instances indicates that the developed approach provides superior quality results and the computational efficiency is demonstrated.