AgentSpeak (L) based testing of autonomous agents

Abstract

Autonomous agents perform on behalf of the user to achieve defined goals or objective. Autonomous agents are often programmed using AgentSpeak language. This language is rich enough to provide necessary support to gain proper functionality within certain environment. Testing of agents programmed in AgentSpeak language is a challenging task. In this paper testing of agents programmed in AgentSpeak has been proposed by deriving AgentSpeak code into goal-plan diagram. Certain coverage criteria have been defined based on the goal-plan diagram. Test cases meeting the defined coverage criteria are used to test the AgentSpeak program with respect to expected functionality. © 2011 Springer-Verlag.