Environment orientation

In an agent based simulation we typically model real world constructs as the entities in a simulation. We are usually most interested in the interaction between these entities as it is from that interaction that the complex behaviours and emergent properties appear.

In this topic we are investigating the notion that agents do not communicate directly with each other. For example, if we imagine that we are simulating agents communicating by sound in the real world then instead of saying that one agent communicates with another we observe that in reality one agent merely impresses its communication upon the environment and other agents take notice of that notion in the environment, or not, as they will.

In this way we decouple the notions of agent to agent communication to one of interaction between an agent and its environment; agents place things in the environment and observe the consequences in the environment of other agents' activities.

In this area of research we are investigating this notion as a way of structuring the interaction of agents in an agent-based simulation of some complex system. We are investigating the consequences of agents getting partial information from their environment and how they combine, and fuse, information from multiple environments whose characteristics vary. For example, visual and olfactory environments have wildly different characteristics but an agent might well use both to determine its future behaviour.

Environment Orientation: An Architecture for Simulating Complex Systems (PDF) Tim Hoverd and Susan Stepney, In Susan Stepney, Peter H. Welch, Paul S. Andrews and Jon Timmis (eds.), Proceedings of the 2009 Workshop on Complex Systems Modelling and Simulation, York, UK, August 2009, pp. 67-82, Luniver Press

A Transactional Architecture for Simulation (PDF) Tim Hoverd and Adam T. Sampson, In ICECCS 2010: Fifteenth IEEE International Conference on Engineering of Complex Computer Systems, pp. 286-290, IEEE Press