Modelling Logistics Chains for Optimal Supply and Maximum Throughput

摘要

This report presents an analysis of a generalised logistics supply chain without back orders. Two methods are proposed: a responsive but robust delivery system based on maintaining set holdings levels; and a technique from Control Theory which pushes stock through the chain in anticipation of demand over both time and space. Furthermore, a heuristic is proposed to set the policy for holdings levels using a hybrid of statistical analysis, Simulated Annealing and Lagrangian Relaxation. Finally, a comparison between methods under uncertainty, with error in prediction and correlated demand, is conducted. Each method was found to be useful in different contexts. The impact of uncertainty and correlated consumption was quantified for a set scenario and both were found to be significant factors in the performance of the supply chain.