Developing A Petri net based Approach to Analyze Agent Systems Feasibility

碩士 === 國立臺灣科技大學 === 工業管理系 === 93 === Agent application is more and more popular, and Agent software system becomes more complex. The development of agent system is usually constructed by Unified Modeling Language (UML), but UML lack dynamic simulation function. Above all, when Agent system is too co...

Full description

Bibliographic Details
Main Authors: Chun-yi Chen, 陳俊乙
Other Authors: Ou-yang Chao
Format: Others
Language:zh-TW
Published: 2005
Online Access:http://ndltd.ncl.edu.tw/handle/15487212151120372680
Description
Summary:碩士 === 國立臺灣科技大學 === 工業管理系 === 93 === Agent application is more and more popular, and Agent software system becomes more complex. The development of agent system is usually constructed by Unified Modeling Language (UML), but UML lack dynamic simulation function. Above all, when Agent system is too complex, constructer can not aware of mistake of Agent system. This research develops a Petri net based approach to analysis Agent systems feasibility. There are three phases in this research. In the concept phase, two kinds of situations, which might affect the feasibilities of an agent system, will be defined. They are: agent roles specification and agent design. The situation of agent roles specification contains: (1)inter-agent interactions and agnet-environment interaction, (2)agent communication. The situation of agent design contains five kinds of agent characteristic: (1)Actuation, (2)Pro-activeness, (3)Deliberative, (4)Communication, (5)Autonomy. In the analysis phase, we interpret how to translate UML into Petri net, and how to simulate and analysis two kinds of situations by Petri net. Finally, in the implement phase, the developed concepts will be implemented in terms of a PDM/ERP collaboration system. That is, a developed UML model for an agent-based PDM/ERP coloration system will be transferred to related Petri net model. The model will be analyzed to find the infeasible situations occurred during PDM/ERP collaboration.