Controling Contract Net Protocol by Local Observation for Large-Scale Multi-Agent Systems

Authors:
Toshiharu Sugawara;Toshio Hirotsu;Satoshi Kurihara;Kensuke Fukuda
Affiliations:
Dept. of Computer Science and Engineering, Waseda University, Tokyo, Japan 1698555;Dept. of Information and Computer Sciences, Toyohashi University of Technology,;Inst. of Scientific and Industrial Research, Osaka University,;National Institute of Informatics, , Chiyoda, Tokyo, Japan 100-000
Venue:
CIA '08 Proceedings of the 12th international workshop on Cooperative Information Agents XII
Year:
2008

Citing 8
Cited 1

The eager bidder problem: a fundamental problem of DAI and selected solutions

Proceedings of the first international joint conference on Autonomous agents and multiagent systems: part 2
The Evolution of the Contract Net Protocol

WAIM '01 Proceedings of the Second International Conference on Advances in Web-Age Information Management
Total performance by local agent selection strategies in multi-agent systems

AAMAS '06 Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems
Gradient field-based task assignment in an AGV transportation system

AAMAS '06 Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems
The Contract Net Protocol: High-Level Communication and Control in a Distributed Problem Solver

IEEE Transactions on Computers
Augmented experiment: participatory design with multiagent simulation

IJCAI'07 Proceedings of the 20th international joint conference on Artifical intelligence
Users matter: a multi-agent systems model of high performance computing cluster users

MABS'04 Proceedings of the 2004 international conference on Multi-Agent and Multi-Agent-Based Simulation
Analysis of learning types in an artificial market

MABS'04 Proceedings of the 2004 international conference on Multi-Agent and Multi-Agent-Based Simulation

Effect of alternative distributed task allocation strategy based on local observations in contract net protocol

PRIMA'10 Proceedings of the 13th international conference on Principles and Practice of Multi-Agent Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

We describe a new adaptive manager-side control policy for the contract net protocol that uses the capabilities of all agents in a massively multi-agent system (MMAS). Recent advances in Internet services, pervasive computing, and grid computing require sophisticated MAS technologies to effectively use the large amount of invested computing resources. To improve overall performance, tasks must be allocated to appropriate agents, and from this viewpoint, a number of negotiation protocols were proposed in the MAS context. Most assume a small-scale, unbusy environment, however. We previously reported the possibility that, using contract net protocol (CNP), the overall efficiency improved by an adequate control of degree of fluctuation in the awarding phase, when the MMAS is in specific states. In this paper, we propose the method to estimate these specific states from the bid values, which have hitherto not been used effectively. Then the new manager-side policy flexibly and autonomously introduces some degree of fluctuation responsive to the estimated states. We also demonstrate that our proposed CNP policy provides considerably better performance than naive CNP and CNP with inflexible policies, even though our policy does not use global information.