An improved algorithm for solving the economic lot size problem (ELSP)

A bounded enumeration combining non-linear optimization, heuristic procedure, and mixed integer programming is proposed to solve the capacitated multi-item lot scheduling problem. The combined effect is demonstrated by a better solution to problem data which has been the subject of inquiry and investigation for over 20 years. In addition, the approach avoids the escape from infeasibility deficiencies of prior solution procedures. The attractiveness of the solution found can be confirmed by the close proximity of the production run frequency to that obtained from the independent solution. Since the independent solution reflects the optimum economic tradeoff between incurring setup costs and inventory carrying costs for each product individually, the advantages of deriving a multi-product production schedule where the frequency of setups conform to these cost relationships are obvious.