Adaptive task resources allocation in multi-agent systems
Proceedings of the fifth international conference on Autonomous agents
Flexible problem-solving roles for autonomous agents
Integrated Computer-Aided Engineering
Automated adaptations to dynamic software architectures by using autonomous agents
Engineering Applications of Artificial Intelligence
A classification framework of adaptation in multi-agent systems
CIA'06 Proceedings of the 10th international conference on Cooperative Information Agents
| Hi-index | 0.00 |
Several formalisms for implementing organizational policies that assign specific roles to each of the agents in a MAS have been proposed like the contract net protocol, the social reasoning mecahnism and the distributed computational economy. However as no single organization is appropriate in all situations, organization self design has been proposed to allow an organization of problem solvers to adapt itself to a dynamically changing environment. In this paper we propose an adaptive organizational framework that exploits the negotiation mechanism of the contract net protocol, social laws of the social reasoning mechanism, market economy feature of the distributed computational economy and organization self design and integrates these features to not only meet the needs of time constrained non time critical applications where computational load on the organization cannot be predicted ahead of time but also to utilize the available system resources efficiently. We also introduce the notion of 'decommitments' and 'penalties' in this approach which was not present in the original contract net protocol.