Ph.D. Priliminary Exam
Topic
Algorithms For Procurement In Transportation Networks
Abstract
With the advent of the Internet, companies are getting interested in e-procurement to reduce cost and time of the procurement operations. In this research work, we address the procurement problem in the context of the transportation networks. We consider two types of network procurement problems. For some industries, e.g., oil, electricity or natural gas supply chain, links of the networks (namely pipelines, transmission lines) are the object of procurement. We call this scenario as edge procurement, because edges are procured to facilitate formation of a network. For other types of businesses, there is a predefined underlying network specified by the sources and destinations of transportation of goods or services. A route in such a network is a set of edges leading from the source node to the destination node. For these industries, the procurement process includes selection of transportation agents and routes. We refer to this as transportation service procurement.
In the context of edge procurement, our research considers a distribution network model where each edge is owned by a selfish owner and a customer wants to supply a product or a service to each of the nodes in the network using a set of edges that results in the lowest possible distribution cost, i.e., a minimum spanning supply tree. The owners incur costs when their edges are used by the customer, and this cost information is privately known only to respective edge owners. We characterize the equilibrium price space of such an economy and analyze a descending price auction mechanism that discovers one of the Walrasian equilibrium prices satisfying certain conditions. We then study the strategic behavior of the owners in such an auction and show that following a greedy strategy is a Nash equilibrium for the owners.
Transportation service procurement auctions are typically characterized by limited number of rounds and low profit margin. In this context, our research addresses the bid determination problem on behalf of truckload carriers for single-lane and combinatorial bids in the successive rounds of an auction, which maximizes the profit. First, we develop a model which selects the best set of bundles to bid on that will maximize the expected profit of the carrier. Second, we compute a profit factor at each round using the industry average price for a lane or lane bundle from the historical figures and the winning bids for the past round of an auction. We develop two methods to update the estimate of the present market rate of a lane or a lane bundle, and two methods to evaluate a profit factor which will maximize the expected profit of the carrier. We perform numerical experiments to validate our results, and also to determine which strategy combination works best for the carrier.
We plan to continue our research in transportation service procurement auctions from the carrier perspective. In the proposed research, we are planning to relax some of our assumptions to make the situation more realistic. Specifically, we are going to look at cases when a carrier is participating in multiple interdependent auctions and when carriers have a possibility of winning a fraction of the total load of a lane or a lane bundle.
Acknowledgements
I would like to thank my adviser, Professor Dharmaraj Veeramani, for giving me an opportunity to work on this problem, and for his constant motivation, guidance and input. Thanks to Professor Harold J. Steudel, Professor Donald B. Hausch, Professor Robert R. Meyer and Professor Andrew J. Miller for agreeing to be on my thesis committee.
I am grateful to Dr. Rao Panchalavarapu, Dr. Erick Wikum, Dave Cobb and Ryan Weight for suggestions and help while working on transportation service procurement auction. Special thanks to my labmate Dr. Debasis Mishra for helping me out at the initial stages of my research in finding out research problems and making me comfortable with research papers. Thanks are also due to my labmate and research colleague Sricharan Poundarikapuram for all the discussions and constructive suggestions.
Finally, a special thanks to my younger sister Shreya, my parents Sadhan and Snigdha and my friend Rashmi for their love and support while pursuing my dreams. This has been a remarkable research journey so far.
Full Document
Full document will soon be available here as pdf.
Final presentation will soon be available here as pdf.
Please contact me by email if you need these documents in the meantime.
|
|
