A branch-and-bound approach for machine selection in just-in-time manufacturing systems

Equipment selection issues are very important in the early stages of implementation of just-in-time (JIT) manufacturing systems. This paper addresses the problem of determining the number of machines for each stage of a JIT system by minimizing production, imbalance and investment costs. The problem is modelled as a mixed-integer nonlinear optimization program and a branch-and-bound algorithm is developed for its solution. This algorithm guarantees the global optimum of the problem and is enhanced by simple, yet very effective, upper bounding heuristics. The solutions obtained by the developed branch-and-bound approach are compared to solutions that have appeared in the literature using heuristic approaches. The comparisons indicate that the proposed algorithm leads to significant economic savings, averaging 17% on a set of problems from the literature. The paper also considers the application of the algorithm to large-scale, industrially-relevant, problems with up to 10 stages and 200 products. Even for the largest of these problems, the search for the integer optimum requires modest computational times. This demonstrates the potential practical impact of the proposed methodology.