Extrospection: agents reasoning about the environment

Current practices for agent-infrastructure interaction enforce agent designers to hard code the name and the use of the infrastructure components in the agent. Therefore, agents can only function within an environment which is a priori known to the agent designer. Even in these environments, agents are not robust against the modification or failure of the infrastructures that may occur at run time. This makes agents fragile in the context of distributed, large scale, and complex environments. We argue that semantic technologies can be employed to overcome these limitations. This paper introduces an ontology called OWL-T that supports the specifications of infrastructure components that allow agents to reason about the functionality and the use at run time. We use the term extrospection for the reasoning done by the agents for the discovery and the use of the infrastructure components based on the tasks and the goals of the agent. A proof of concept implementation illustrates the use of OWL-T in a multi-agent foraging scenario.