Agent-based resource discovery architecture for environmental emergency management

Abstract

An agent-based environmental emergency management framework is introduced as a loosely coupled collection of agents that can cooperate to prepare for and response to environmental emergency situations. In this framework, resources play a critical role because they are the foundation for taking action in environmental emergencies. Therefore, an agent-based resource discovery architecture is then proposed to search for the relevant resources over the Internet. In the making of an agent-based resource discovery architecture, two pivotal issues need to be addressed: resource description language (RDL) and its resource matchmaking mechanism. RDL provides a specification to publish and request for resources in environmental emergency situations, and matchmaking is the process of finding an appropriate resource for a request through a medium. In this paper, a possibilistic Petri net-based resource description language is proposed as an advanced RDL with four key features: possibilistic transitions to represent a resource or a request; input places to denote preconditions expected to hold before performing the resources; output places to denote postconditions expected to hold after performing the resources; possibility and necessity measures to quantify the confidence levels that an agent can provide the relevant resource for a request. A matchmaking mechanism, permitting a relaxed match for close semantics, is also developed to search for the possible resources among agents for a request.