A computational framework for studying decentralized supply chain dynamics
In a typical decentralized supply chain, i.e., one with multiple loci of control, the dependencies between the entities are formalized through contracts. The parameters over which contracts are observed in a typical buyer-seller system include price, periodicity of ordering, quantity commitments, and delivery commitments. The central idea behind studying these parameters of interaction is to choose contract forms and values that are beneficial to both the contracting entities. Representation of supply chain systems at a level sufficient to capture the essential system interactions requires a high-level architecture that combines simulation and optimization techniques. In this framework, the simulator captures the uncertainties while the optimizer captures the combinatorial nature of manufacturing decisions. This computational framework is applied to a two-entity retailer-manufacturer system to investigate the quantity flexibility contract and its role in mitigating the bullwhip effect, an amplification of demand variability that can arise under certain conditions. This amplification can have a significant negative effect on the manufacturer. Through the quantity flexibility contract, the manufacturer aims to reduce the variability in the retailer's orders while leaving the retailer no worse off. In this thesis, we quantify the impact of a particular contract setting on both entities, which consequently allows us to test for its feasibility. Results that illustrate the effect of the contract parameters on both entities are also presented. We further investigate the effect of the manufacturing function on the contract. Through these analyses, we demonstrate the use of the computational framework in studying decentralized supply chain dynamics, and, in particular, the utility of detailed manufacturing models. ^
Major Professors: Joseph F. Pekny, Purdue University, Gintaras V. Reklaitis, Purdue University.
Engineering, Chemical|Engineering, Industrial