Reformulations and solution methods to the integrated lot sizing and scheduling problem

Abstract

Production planning decisions have always had a great influence on the industries performance. Issues such as lot sizing, scheduling, adequacy of inventory levels, market demand and the consequent synchronization of all these decisions are still a challenge. In some industries, if the lot sizing and scheduling decisions are taken independently it may cause difficulties and prevent the flexibility of the production process to attend the market changes. Optimization models that integrate decisions of lot sizing and scheduling has been proposed in the literature using two main approaches related to the scheduling decisions. One approach uses the idea of dividing the periods into subperiods - the GLSP model (General Lot-sizing and Scheduling Problem), the other uses subtour eliminations constraints associated to the traveling salesman problem to obtain the production schedule. These approaches were used for solving the integrated problem in many industrial contexts obtaining good solutions, but without proof of optimality in most cases. The objective of this project is to propose alternative models for the lotscheduling problem exploring various strategies for modeling the scheduling decisions. It also includes the development of solution methods to obtain good solutions in a reasonable computational time. It is expected to obtain models and solution methods with a better computational performance in terms of solution time and/or solution quality compared to other models in the literature.