No Agent is an Island: A Framework for the Study of Inter-Agent Behavior
We describe a framework for the study of inter-agent behavior. Our starting point is the notion that choosing to perform an action will constrain the capacity to choose other actions, both for the agent concerned and for the other agents with which it interacts. We represent these conflicts between choices in an abstract framework using an "option framework". In an option framework we are not concerned with the nature of the actions that can be chosen, only the ways in which they conflict with the other choices in the framework.
We can then partition the options in the framework according to the agent which can select them, and associate utilities, with respect to all the agents in the system, with them. Agents will then select options according to the constraints imposed by conflicts between actions. Where the choices of two agents conflict, the conflict is resolved according to which agent controls the conflict, resulting in the realization of some subseet of the choices. The agent can then evaluate the realized actions according to a function which may take into account the utility produced for itself and other agents. The task of the agent is to select the set of actions which produces the subjectively most favoured realization.
Having formally presented the framework we show how it can be used to explore the inter-agent behavior of systems of agents according to a number of factors which will determine how they go about their task. We also show how some other approaches to the investigation of the inter-agent behavior can be modeled in our framework.
The framework is sufficiently abstract to provide the means to explore all aspects of inter-agent behavior by both empirical and analytic means. We give examples of some hypotheses that may be investigated in this framework.[Full Paper]
For each technical report listed here, copyright and all intellectual property rights remain with the respective authors. Copyright is effective from the year of publication in each case. By downloading a file from this page, you agree to use it only for purposes of research and scholarship. Any other use of this material or storage of it in any medium or its sale or distribution in any form is expressly forbidden without prior written permission from the authors concerned.