GOALI: Supply Chain Optimization under Uncertainty for the Chemical Process Industries

The development of planning models and algorithmic techniques for optimizing the supply chains of chemical process industries will be performed. The goal is to provide tractable models and solution techniques for multisite deterministic and stochastic planning problems of realistic size and industrial relevance. A hierarchical decomposition procedure is proposed based on Lagrangean relaxation. The basic concept is to partition the original problem into smaller, more tractable subproblems through the hierarchical relaxation of key complicating constraints. The systematic identification of these constraints is accomplished by utilizing dual multiplier information. Preliminary results of this framework suggest significant reduction in computation time. The stochastic work will focus on utilizing the deterministic equivalent concepts for providing a computationally tractable description of uncertainty in multisite planning problems. Normally distributed uncertainty on product demands, prices, and production rates will be explored. Emphasis will be placed on answering questions regarding inventory level management, safety stocks, and capacity utilization in the face of uncertainty. The concept of linking the computational study on the deterministic model to the solution of the stochastic problem could have significant impact in industrial planning.