STRUCTURED BUFFER-ALLOCATION PROBLEMS

We study the effect of changing buffer sizes in serial lines with gene ral blocking, a mechanism that incorporates limited intermediate finis hed goods inventory at each stage, as well as limited intermediate raw material inventory. This model includes ordinary manufacturing, commu nication, and kanban blocking as special cases. We present conditions under which increasing buffer sizes or re-allocating buffer capacity i ncreases throughput, and in some cases characterize optimal allocation s. Our comparisons hold pathwise; they depend on structural properties of a line but not its service time distributions. The key to our anal ysis is formulating the model as a generalized semi-Markov process (GS MP). We show that a serial line with general blocking possesses variou s structural properties from which pathwise comparisons follow directl y, making a detailed case-by-case analysis unnecessary These results h elp illustrate the power of GSMPs as a modeling framework.