Open Access
Solo, Christopher James
Graduate Program:
Industrial Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
May 28, 2009
Committee Members:
  • Arunachalam Ravindran, Dissertation Advisor
  • Arunachalam Ravindran, Committee Chair
  • Soundar Rajan Tirupatikumara, Committee Member
  • M Jeya Chandra, Committee Member
  • Susan H Xu, Committee Member
  • supply chain
  • uncertainty
  • stochastic optimization
  • robust optimization
  • chance-constrained goal programming
This research involves the development of a flexible, multi-objective optimization tool for use by supply chain managers in the design and operation of manufacturing-distribution networks under uncertain demand conditions. The problem under consideration consists of determining the supply chain infrastructure; raw material purchases, shipments, and inventories; and finished product production quantities, inventories, and shipments needed to achieve maximum profit while fulfilling demand and minimizing supply chain response time. The development of the two-phase mathematical model parallels the supply chain planning process through the formulation of a strategic submodel for infrastructure design followed by a tactical submodel for operational planning. The deterministic strategic submodel, formulated as a multi-period, mixed integer linear programming model, considers an aggregate production planning problem in which long-term decisions such as plant construction, production capacities, and critical raw material supplier selections are optimized. These decisions are then used as inputs in the operational planning portion of the problem. The deterministic tactical submodel, formulated as a multi-period, mixed integer linear goal programming model, uses higher resolution demand and cost data, newly acquired transit time information, and the previously developed infrastructure to determine optimal non-critical raw material supplier selections; revised purchasing, production, inventory, and shipment quantities; and an optimal profit figure. The supply chain scenario is then modified to consider uncertain, long-term demand forecasts in the form of discrete economic scenarios. In this case, a multi-period, mixed integer robust optimization formulation of the strategic submodel is presented to account for the probabilistic demand data. Once the stochastic strategic submodel is presented, short-term, uncertain demand data is assumed to be available in the form of continuous probability distributions. By modifying decision makers’ objectives regarding demand satisfaction, the distribution-based demand data is accounted for through the development of a multi-period, mixed integer chance-constrained goal programming formulation of the tactical submodel. In order to demonstrate the flexibility of both the deterministic and stochastic versions of the overall two-phase model, numerical examples are presented and solved. The resulting work provides supply chain managers with a flexible tool that can aid in the design and operation of real-world production-distribution networks, where uncertain demand data is available at different times and in various forms.