Reducing complexity of autonomous control agents for verifiability

Abstract: The AgentSpeak type of languages are considered for decision making in autonomous control systems. To reduce the complexity and increase the verifiability of decision making, a limited instruction set agent (LISA) is introduced. The new decision method is structurally simpler than its predecessors and easily lends itself to both design time and runtime verification methods. The process of converting a control agent in LISA into a model in a probabilistic model checker is described. Due to the practical complexity of design time verification the feasibility of runtime probabilistic verification is investigated and illustrated in the LISA agent programming system for verifying symbolic plans of the agent using a probabilistic model checker.