An improved lower bound for the changeover scheduling problem

In a multi-product, discrete-item manufacturing environment, it is often more economical to produce several products on one flexible machine than to have a dedicated machine for each product. In this paper, we consider the changeover scheduling problem (CSP). The CSP minimizes the cost consequences of changing over limited-capacity resources when two or more products are produced on them. We consider scheduling multiple products on a single machine where demand is dynamic, deterministic, and is to be met without backlogging. We develop analytical results for a lower bound on this problem and give computational results using this lower bound in a branch-and-bound algorithm.