An Agent-based Commodity Trading Simulation

• Main Authors: CHENG, Shih-Fen, LIM, Yee Pin, LIU, Chao-Chi
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2009
• Subjects: Computational Economics; Commodity Trading; Artificial Intelligence and Robotics; Business; Operations Research, Systems Engineering and Industrial Engineering
• Online Access: https://ink.library.smu.edu.sg/sis_research/839; https://ink.library.smu.edu.sg/context/sis_research/article/1838/viewcontent/aamas09_commodity_demo.pdf
• Institution: Singapore Management University

In recent years, the study of trading in electronic markets has received significant amount of attention, particularly in the areas of artificial intelligence and electronic commerce. With increasingly sophisticated technologies being applied in analyzing information and making decisions, fully autonomous software agents are expected to take up significant roles in many important fields. This trend is most obvious in the financial domain, where speed of reaction is highly valued and significant investments have been made in information and communication technologies.Despite the successes of automated trading in many important classes of financial markets, commodity trading has lagged behind, mainly because of its complicated product categorization and logistical fulfillment considerations. These two factors greatly hinder automation efforts because whenever an event that has significant physical impact on the commodity supply chain occurs, complicated and commodity specific reactions (might include trading, re-hedging, or even logistic adjustment, to name just a few) would be required. Due to this reason, to master even just a particular commodity market would take several years of intensive training and exposure. To facilitate better understanding on the event-centric commodity market, we built an agent-based commodity trading simulation that is driven by physical events [1]. The simulation platform serves two purposes: First, it is used as a tool that allows more effective training; second, professional trader's behaviors in face of uncertain events could be measured comprehensively for thorough analysis.