Design and Implementation of Service Discovery Architecture Based on Multi-Agent Systems in an Ad-Hoc Environment - For the Observing and Recording System

Since the mobile device characters features in motion while traditional information services are unable to meet demands of the mobile environment, this study therefore proposes a “multi-agent architecture – a DASS (Distributed Agent-based Service Sharing) –with the purpose to design and implement a service sharing architecture using the autonomy and communication abilities of agents, to actively discover services to mediate among shared services, so as to allocate and manage the distributed service. The system development process follows PASSI (a Process for Agent Societies Specification and Implementation) methodology to form “multi-agent society” models. System development of this study is carried out on the agent middleware platform, JADE (Java Agent DEvelopment Framework); substantial contributions are: (1) Establishment of service-sharing System by using the agent technology that uses a single window for communicating with other agents; (2) the system framework and the service discovery template accomplished in this study enable the developers to promptly configure service-sharing applications; (3) inheriting the abstractive service search mechanism of FIPA (The Foundation for Intelligent Physical Agents), this System has increased compatibilities with other systems; and using the existing JXTA based service discovery architecture to implement service distribution and search on the agent platform resolves difficulties in accessing peer to peer services in the Ad-Hoc network; (4) the System sets up mediation models for service-sharing according to agent communication language set up with (defined by) FIPA standards, and sets up communication ontology for service-sharing, therefore in the communication process, it is easy to understand intentions of each other without the need of particular understanding of the vocabulary and syntax used by each other; this substantially simplifies communication problems between heterogeneous agents.