The part-period algorithm and least-total-cost-heuristics in single-level discrete-demand lot sizing

The most misunderstood of all heuristics used in the single-level discrete-demand lot-sizing environment are the least total cost and the part-period algorithms. Which of these heuristics is better suited in this case depends upon the method used to calculate holding costs. Applying both heuristics to the same data set, it appears that the past-period algorithm using end-of-period holding costs provides the best results.