Multi agent simulation of artificial foreign exchange market

• Main Author: Maria Myrna Handoko.
• Other Authors: Ong Yew Soon
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Computer science and engineering::Computer applications::Computers in other systems
• Online Access: http://hdl.handle.net/10356/17035
• Institution: Nanyang Technological University

Foreign exchange market is one of the largest and most liquid financial markets in the world. Traders trade 24 hours non-stop and it greatly affects foreign exchange rate volatility. In this study, a multi-agent system is built to simulate an artificial foreign exchange market approach to analyse rate movements for a certain period. It helps traders and analysts to predict, make the correct decision, and earn big profits. Other than that, few psychological behaviors of the market are incorporated and discussed as part of the simulation results. This model of artificial foreign exchange market adopted the AGEDASI TOF framework, a Genetic-Algorithmic Double Auction Market Simulation in Tokyo Foreign Exchange Market, by Izumi 1998. The simulation model is constructed and presented using Java-Agent Development (JADE) framework in collaboration with Evolutionary Computation Java-based (ECJ) library to accommodate the interaction of agents’ learning behavior with genetic algorithms. This learning method is to improve traders’ prediction by replacing their opinion with others who predict more accurately. Finally, the emergent phenomena are to be analyzed based on the external and internal data used for simulation inputs. External data include macroeconomic news and fundamentals; and internal data represent the market movements within the simulation period. Both inputs are particularly described and considered throughout the decision making process of traders.