We use agent-based models in our research to help us gain insight into the collective behaviour of people. In such models each agent has its own goals and is capable of perceiving its environment, making decisions and performing actions independently.
The fully automated trading system we have developed for Rudyerd is a multi-agent system that works with artificial traders each with its own view of the market and its own rationale for making trading decisions. The system is used to trade automatically in the foreign exchange market.The pedestrian simulator we have developed for Legion is used to build models with large numbers of individual agents (simulated people). The models are used to predict the crowd behaviour that is likely to result from the interactions of the agents in a particular environment.