Semantic-based dynamic positioning mechanism for problem solving in multi-agent systems
来源期刊：中南大学学报(英文版)2014年第2期
论文作者：LI Qing-shan(李青山) CHU Hua(褚华) XUE Bao-ye(薛宝叶) ZHANG Chao(张超)
文章页码：618 - 628
Key words：multi-agent systems (MAS); task revolution; semantics; dynamic location
Abstract: In multi-agent systems (MAS), finding agents which are able to service properly in an open and dynamic environment are the key issue in problem solving. However, it is difficult to find agent resources quickly and position agents accurately and complete the system integration by the keyword matching method, due to the lack of clear semantic information of the classical agent model. An semantic-based agent dynamic positioning mechanism was proposed to assist in the system dynamic integration. According to the semantic agent model and the description method, a two-stage process including the domain positioning stage and the service semantic matching positioning stage, was discussed. With this mechanism, proper agents that provide appropriate service to assign sub-tasks for task completion can be found quickly and accurately. Finally, the effectiveness of the positioning mechanism was validated through the in-depth performance analysis in the application of simulation experiments to the system dynamic integration.
LI Qing-shan(李青山), CHU Hua(褚华), XUE Bao-ye(薛宝叶), ZHANG Chao(张超)
(Software Engineering Institute, Xidian University, Xi’an 710071, China)
Abstract:In multi-agent systems (MAS), finding agents which are able to service properly in an open and dynamic environment are the key issue in problem solving. However, it is difficult to find agent resources quickly and position agents accurately and complete the system integration by the keyword matching method, due to the lack of clear semantic information of the classical agent model. An semantic-based agent dynamic positioning mechanism was proposed to assist in the system dynamic integration. According to the semantic agent model and the description method, a two-stage process including the domain positioning stage and the service semantic matching positioning stage, was discussed. With this mechanism, proper agents that provide appropriate service to assign sub-tasks for task completion can be found quickly and accurately. Finally, the effectiveness of the positioning mechanism was validated through the in-depth performance analysis in the application of simulation experiments to the system dynamic integration.
Key words:multi-agent systems (MAS); task revolution; semantics; dynamic location
