CoMMA-COGs

The CoMMA-COGs project (Cooperative Man Machine Architectures - Cognitive Architecture for Social Agents) is part of the Multi-Agent Systems group's larger aim to develop integrated architectures for multi-agent systems and thus to counteract the current tendency in Artificial Intelligence research to concentrate on improving competence in ever more specific areas. In this larger context we hope to give an account of motivation in integrated agent architectures that exploits relevant research in Cognitive Science.

A new branch of computer applications is opening up based on virtual animated worlds, where human users, their software representatives and a host of services co-exist in a networked environment. This promises to be a major area for the application of agent-oriented techniques. However, these applications demand multi-agent systems that are more ,exible than current systems and integrate a wide variety of functionality in a dynamic fashion. The key to achieving this integration is to take seriously the notion of bounded rationality and develop an architecture which manages the resources of an agent consistent with changes in the environment. It should also share resources out to sub-components of the agent according to the way they contribute to the agent's overall goals. The agent society, correspondingly, must be structured to take into account the resource managing behaviour of the individual agents.

The COGs project has three major technical goals:

• We will build on the tradition of multi-agent research in the group to design resource-aware agent societies. In doing so we will investigate cooperation protocols that embody fair mechanisms for sharing resources and incorporate constructs into the architecture that will support the dynamic specification, monitoring and policing of accepted social laws.
• We will develop a framework for self-organisation in agent societies, which allows individual agents to aggregate dynamically with other agents to form a larger entity which itself represents an agent to the outside environment (a holon). In particular, we will look at the impact of social laws on the process of self-organisation and investigate how particular organisational structures maximise resource utilisation.
• We will develop a resource-aware agent architecture that allows agents to flexibly manage their resources, whilst pursuing multiple goals in an unpredictable environment. We will produce a general model of abstract resources and introduce constructs into the architecture that monitor resource usage and help to adapt it to a changing environment. We will pay particular attention to the way this aspects the flow of control within our current agent architecture and will compare the resulting architecture with models documented in the Cognitive Science literature.

The results of the project will be of use in a variety of application areas. They will provide a more flexible agent architecture for the current main applicationarea of the group: dynamic scheduling in haulage firms. This application area will benefit particularly from the new techniques for dynamic reorganisation of agent societies. In addition, we will develop a toolkit for simulating societies of agents and thus exercise the full range of the technology we are developing. The toolkit will be based on animated objects in distributed, virtual worlds. The group will also test the toolkit in two other areas. First, it will be used to provide an intuitive, interactive environment for Internet sites ordering electronic commerce. The work in COGs will be particularly useful in supporting the resource-constrained aspects of the bargaining process, and in developing constructs for interaction that foster fairness and trust. Second, we will investigate the mentalistic interpretation of the resource-bounded constructs to support the representation of lifelike characters in entertainment software such as interactive drama and games.

For more information contact Alastair Burt

Last change: September 30, 1998