Knowledge Base Support for Design and Synthesis of Multi-Agent Systems
Versandkostenfrei!
Versandfertig in über 4 Wochen
54,99 €
inkl. MwSt.
PAYBACK Punkte
27 °P sammeln!
Agent Tool is an AFIT-produced, AFOSR-sponsored multi-agent system (MAS) development tool intended for production of MASs that meet military requirements. This research focuses on enabling MAS design and synthesis tools like agent Tool to store, retrieve, and filter persistent, reusable, and reliable agent domain knowledge. This "enabling" is vital if such tools are expected to produce consistent, maintainable, and verifiable agent applications on short timetables. Enabling requires: 1) modeling the agent knowledge domain, 2) designing and employing a persistent knowledge base, and 3) bridging...
Agent Tool is an AFIT-produced, AFOSR-sponsored multi-agent system (MAS) development tool intended for production of MASs that meet military requirements. This research focuses on enabling MAS design and synthesis tools like agent Tool to store, retrieve, and filter persistent, reusable, and reliable agent domain knowledge. This "enabling" is vital if such tools are expected to produce consistent, maintainable, and verifiable agent applications on short timetables. Enabling requires: 1) modeling the agent knowledge domain, 2) designing and employing a persistent knowledge base, and 3) bridging that domain model to the knowledgebase with an extensible domain interchange grammar. The achieved interchange grammar, called Multi-Agent Markup Language (MAML), is presented and shown to be capable of representing MAS design knowledge in a concise and easily parsed form that is readily stored and retrieved in the knowledge base. The selected knowledge base, called the Agent Random-Access Meta-Structure (ARAMS), is shown to support MAML and operate in a distributed environment that permits sharing of agent development knowledge between various tools and tool instances. Tests of MAML and ARAMS with agent Tool are summarized, and related future work suggested.
Andere Kunden interessierten sich für
