给出了一个多Agent松散耦合的网格服务发现框架.在该框架中把Agent分为三类:服务Agent、需求Agent和服务发现Agent.提出以模糊Petri网作为服务描述语言的规范,用以发布或请求服务,用可能性变迁表示一个服务或一个需求,输入库所代表在提供(或请求)服务前需要成立的前提条件,输出库所表示提供(或获得)服务后成立的条件,用可能性与必然性来量化一个服务Agent能对一个请求提供相关服务的信心程度.最后基于本体论给出了一个支持部分匹配的服务匹配算法,并用一个车辆维修服务系统的例子进行了说明.
A multi-agent-based grid service discovery framework was put forward, in which agents were distinguished into three categories: service-agent, request-agent and service-discovery-agent. A service description language based on fuzzy Petri nets (FPN-SDL) was proposed as a specification to publish or request for a service, it used possibility transition to represent a service or request, input places and output places to denote conditions before and after performing the service respectively, and possibility and necessity to quantify the confidence levels that the service could satisfy a request. At last this paper presented a semantic-based service matchmaking that can find an appropriate service for a request and support partial matching, and illuminated these through an example of the vehicle maintenance service system.