PROGRAMME SCHEDULE

09:00

Registration

09:30

Welcome
Prof. Steve Miller
Dean, School of Information Systems
Singapore Management University

Opening
Dr. Robert De Souza
Executive Director, TLI – Asia Pacific

09:45

Decision Support in Chemical Supply Chains
Prof. I.A. Karimi
National University of Singapore

Chemical cluster is the second largest contributor to Singapore's manufacturing output. The chemical supply chains have their own distinct and special characteristics and we have been addressing their needs for the last several years. In this talk, we will give an overview of our research activities in chemical supply chain management and chemical logistics. We will specifically describe how agent-based methodologies have been used in modeling and simulating petroleum supply chains, and they, along with other technologies such as optimization and simulation, provide a comprehensive umbrella framework for integrated decision support on several issues such as supply chain design, policy evaluation, disruption management, etc.
 

10:15

Multi-Period Combinatorial Auctions for Decentralized Resource Allocation
Associate Prof. Lau Hoong Chuin
Singapore Management University

In this talk, we consider a resource allocation and scheduling problem where information and management decisions are decentralized. This problem is motivated by a large-scale dynamic resource allocation problem in a megascale container terminal. We propose a generic combinatorial auction mechanism that allows agents to competitively bid for the resources needed in a multi-period setting, regardless of the respective scheduling problem faced by the agent. We then show optimized bidding and price adjustment strategies can be designed to solve the large-scale resource coordination problem in the container terminal.
 

10:45

Refreshments

11:00

Designing a Market Game for the Global Commodity Supply Chain
Assistant Prof. Cheng Shih-fen
Singapore Management University

Closer vertical integration has become prevalent in many agri-food commodity supply chains. This new paradigm has promoted international traders to expand their roles to more than just trading, but also many physical activities. As a result, understanding the risks embedded in these complicated activities and also hedging these risks properly has become more and more important. In this talk, I present a market game that can be used in modeling and studying risk associated with trading, and propose how other physical activities could be included as extensions. A pilot experiment focusing on the trading risk was run with a mixture of human traders and an autonomous agent. With this experiment, we demonstrate that market games are indeed quite effective in promoting human traders to learn from the simulated rare events as well as extracting the patterns of those successful human traders.
 

11:30

A Supply Chain Management Trading Agent Competition
Prof. Michael Wellman
University of Michigan

For the past eight years, autonomous software bidding agents have pitted their skills against each other in several market games as part of an annual international Trading Agent Competition (TAC) event. The TAC Supply Chain Management (TAC/SCM) game, introduced in 2003, presents a particularly challenging trading scenario, where agents representing PC manufacturers compete for customers and negotiate with suppliers in order to generate profits over a simulated production year. Over the years, TAC/SCM participants have used this domain to develop and evaluate new ideas in factory scheduling under uncertainty, price prediction, decision-making architectures, and strategic analysis, among other relevant subjects. Michigan's entry, Deep Maize, has introduced several new techniques, and enjoyed relative success in the tournament. I present examples from the design and analysis of Deep Maize, and highlight several key lessons from our TAC/SCM experience.
 

12:30

End of Seminar. Networking Luncheon