An Auction-Theoretic Modeling of Production Scheduling to Achieve Distributed Decision Making Dissertation Proposal

Most existing methods for scheduling are based on centralized or hierarchical decision making using monolithic models. In this study, we investigate a new generation of scheduling methods based on a distributed and locally autonomous decision structure. Speciically, we propose a decision structure based on auction theory. The basic idea is to localize and distribute the functionality of scheduling, leaving the complexity of operational decisions to local decision makers, while maintaining a simple and generic central coordination mechanism. The proposed structure allows local decision makers to make their decisions dynamically and independently according to changes in their local environments. A central coordination mechanism then makes resource allocation based on an iterative auction process using information obtained from local decision makers. We propose following research endeavors: We will study the decomposition of monolithic optimization models which provides the basis for a distributed decision structure. In this research, we will focus on combinatorial optimization models in job shop scheduling. Speciically, we propose to study Lagrangean relaxation and decomposition techniques. We will study auction theoretic modeling of production scheduling which provides the means for analyzing how local preferences might be aggregated into collective decisions. Speciically, we propose to investigate two types of auction mechanisms: (1) price-adjustment based auction, and (2) winner coalition based auction. To evaluate practicality of the proposed distributed structure, we plan to test our approach using real data and operational scenarios at Ford North Penn Electronics Facility.