2002 Fiscal Year Final Research Report Summary
Analysis and Control of Dynamics of Multiagent System
| Project/Area Number |
12680365
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Intelligent informatics
|
|---|
| Research Institution | Tohoku University |
|---|
Principal Investigator |
KINOSHITA Tetsuo Tohoku University, Information Synergy Center, Professor, 情報シナジーセンター, 教授 (20282006)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
AGU Masahiro Ibaraki University, Faculty of Engineering, Professor, 工学部, 教授 (10016462)
KATO Takashi Tohoku University, Research Institute of Electric Communication, Research Assistant, 電気通信研究所, 助手 (20323115)
SUGANUMA Takuo Tohoku University, Research Institute of Electric Communication, Associate Professor, 電気通信研究所, 助教授 (70292290)
|
|---|
| Project Period (FY) |
2000 – 2002
|
| Keywords | Agent / Multiagent / Artificial Intelligence / Distributed Artificial Intelligence / Network |
|---|
| Research Abstract |
Mutiagent System consists of many agents which can act autonomously and cooperatively. The behavioral properties and functions of a multiagent system as a whole should be determined by various mutual actions among agents together with the behavioral properties of each agent. Since the behavior of multiagent system should also be non-deterministic and situation-dependent, the designers of such system cannot describe and specify the detailed behavior of the system nor exclude the illegal actions from the system.
In this research project, we studied the methods to analyze and control the behavioral properties of multiagent system in order to get knowledge on the design and development of multiagent systems. As a result, the following new knowledge can be acquired, i.e.,
(a) To improve the operation to form the multiagent organization based on the task assignment actions, the agents have to have and utilize information on other agents which can work together.
(b) In the multiagent systems in which each agent has to determines its actions by receiving information on resources, suitable delay for receiving information can improve the performance of the system.
(c) In a multiagent application for mutimedia communication services, we can design and implement the agents which are responsible to observe and control the runtime properties of the application. By using such agents, we can build the multiagent systems which can maintain the required QoS (Quality of Service) against various changes of behavioral environment of the application.
|
