Modelling and verifying BDI agents with bigraphs

Abstract

• We show BDI agent programming languages can be encoded in Milner's Bigraphs. • We prove the encoding of a superset of AgentSpeak, e.g. CAN language, is faithful. • We provide a verified executable semantics for CAN language programs. • We verify CAN language program properties through model checking. • This is a foundation to enable advanced reasoning, e.g. probabilistic verification. The Belief-Desire-Intention (BDI) architecture is a popular framework for rational agents; existing verification approaches either directly encode simplified (e.g. lacking features like failure recovery) BDI languages into existing verification frameworks (e.g. Promela), or reason about specific BDI language implementations . We take an alternative approach and employ Milner's bigraphs as a modelling framework for a fully featured BDI language, the Conceptual Agent Notation (CAN)—a superset of AgentSpeak featuring declarative goals, concurrency, and failure recovery. We provide an encoding of the syntax and semantics of Can agents, and give a rigorous proof that the encoding is faithful. Verification is based on the use of mainstream software tools including BigraphER, and a small case study verifying several properties of Unmanned Aerial Vehicles (UAVs) illustrates the framework in action. The executable framework is a foundational step that will enable more advanced reasoning such as plan preference, intention priorities and trade-offs, and interactions with an environment under uncertainty.