A Multi-Criteria Optimization Model for a Four-Stage Multi-Product Supply Chain

Open Access
Niman, Emily
Graduate Program:
Industrial Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
Committee Members:
  • Dr Ravindran, Thesis Advisor
  • Arunachalam Ravindran, Thesis Advisor
  • supply chain
  • multi-criteria
  • multicriteria
  • optimization
  • supply chain modeling
  • multi-product
  • supply chain management
  • supply chain efficiency
The goal of this thesis is to develop a supply chain model that integrates several common supply chain modeling techniques into a realistic and comprehensive general model. The model addresses production, transportation, and inventory decisions, which are often considered separately. Supply chain modeling problems tend to be inherently multi-criteria problems in that the optimization of several objectives simultaneously is required in order to develop the most efficient supply chain. In this thesis up to three criteria are considered simultaneously and are used to develop a multi-criteria optimization model. The three criteria are profit, customer service, and inventory capital. The solution to the model can be used as a supply chain management tool to determine the quantities and scheduling related to the production and distribution of products across a supply chain in order to best optimize the objectives. The model developed is for a four-stage, centralized supply chain in which two products are being produced and distributed. There are several suppliers, manufacturers, warehouses, and retailers at each stage and each manufacturer has three production lines. Two modes of transportation are available between each stage of the supply chain and a freight rate function is utilized. The freight rate function follows an All-Units discount cost structure that depends on the weight of shipments and the transportation mode being used. The demand at the retailers drives the supply chain and is deterministic and independent among the retailers. The supply chain is modeled as a mixed integer linear program that is solved using GAMS (General Algebraic Modeling System). The multi-criteria model is solved using goal programming, a multi-criteria modeling technique that aims to minimize the total deviation of the objective values from their goals. The problems are solved either as a preemptive goal program or non-preemptive goal program, depending on how the criteria preference information is specified. An illustrative example is used to demonstrate the implementation of goal programming and the overall use of the model. Two bi-criteria cases are modeled in which profit and customer service are maximized. Additionally, a third case addresses changes in the solution when a third objective, the minimization of inventory capital, is added to the model. The solutions to all three cases are compared and differences among preemptive and non-preemptive solutions are identified, as well as the differences between the bi-criteria and multi-criteria models. The model not only incorporates several supply chain modeling techniques into a single model, it also addresses multiple criteria simultaneously in an attempt to develop a realistic supply chain model that can be used as a tool to improve supply chain efficiency.