Api-calculus for intelligent-agent formal modeling and its application in distributed geospatial data conflation.

摘要

Intelligent-agent paradigm is making a wide range of exciting new distributed applications possible. However, beyond the basic engineering and development challenge in design and implementation of agent-based systems, there are several security, validation, verification, and performance related questions that need to be answered. To address these and many other concerns, it is necessary to utilize formal modeling methods useful for representing such systems. Several calculi have been introduced for this purpose. Polyadic pi-calculus, higher order pi-calculus, ambient calculus, and concurrent object-oriented Petri nets are some of the most popular tools for modeling systems consist of agents. However, none of the above calculi covers all the characteristics of intelligent-agent based systems. Intelligence, natural grouping, security, and migration aspects of such systems are not completely covered by any of the above calculi.; Api-calculus is introduced as an extension to pi-calculus. In this calculus we introduce three core concepts: knowledge unit, milieu and term. These concepts give api-calculus the capability to present intelligence in agent-based systems, using knowledge units , as well as natural grouping of agents, using milieus. It also has the potential for security representation, using milieus and terms.; The main motivation for the development of api-calculus was to be employed for modeling an agent-based distributed geospatial data conflation system. This system is a multi-level intelligent-agent based system for evaluation, collection and integration of geospatial data from different data repositories.; The architecture of the agent-based distributed geospatial data conflation system is presented. Next, the syntax and semantics of api-calculus is introduced. Finally, the apicalculus is employed to model the distributed geospatial data conflation system.