Synthesis of Optimization and Simulation for Multi-Period Supply Chain Planning with Consideration of Risks

Document Type



Doctor of Philosophy


Business Administration

Date of Defense


Graduate Advisor

L. Douglas Smith


Navarro, Virginia

Donald C. Sweeney

Haitao Li

Lorena Bearzotti


Solutions to deterministic optimizing models for supply chains can be very sensitive to the formulation of the objective function and the choice of planning horizon. We illustrate how multi-period optimizing models may be counterproductive if traditional accounting of revenue and costs is performed and planning occurs with too short a planning horizon. We propose a “value added” complement to traditional financial accounting that allows planning to occur with shorter horizons than previously thought necessary. This dissertation presents a simulation model with an embedded optimizer that can help organizations develop strategies that minimize expected costs or maximize expected contributions to profit while maintaining a designated level of service. Plans are developed with a deterministic optimizing model and each of the decisions for the first period in the planning horizon are implemented within the simulator. Random deviations in demands and in upstream and downstream shipping times are imposed and the state of the system is updated at the end of each simulated period of activity. This process continues iteratively for a chosen number of periods (90 days for this research). Multiple replications are performed using unique random number seeds for each replication. The simulation model generates detailed event logs for each period of simulated activity that are used to analyze supply-chain performance and supply-chain risk. Supply-chain performance is measured with eleven key performance indicators that reveal system behavior at the overall supply-chain level, as well as performance related to individual plants, warehouses, and products. There are three key findings from this research. First, a value-added complement in an optimization model’s objective function can allow planning to occur effectively with a significantly shorter horizon than required when traditional accounting of costs and revenues is employed. Second, solutions with the value-added complement are robust for situations where supply-chain disruptions cause unexpected depletions in inventories at production facilities and warehouses. Third, ceteris paribus, the hybrid multi-period planning approach generates solutions with higher service levels for products with greater revenue per average production-minute, shorter average upstream lead times, and lower coefficients of variation for daily demand.

This document is currently not available here.