This paper extends previous work on reducing setup cost in the dynamic
lot-sizing problem with deterministic but time-varying demands. Using
a marginal cost approach, a simple heuristic is developed to find the
optimal production schedule for the reduced setup cost and the optima
l investment in setup reduction. A numerical example is solved for the
declining linear and exponential setup cost reduction functions, whic
h respectively yield a concave and a convex total cost. Furthermore, a
n extensive experiment is performed to evaluate the efficiency of the
heuristic, which is found to perform well in certain operational setti
ngs.