No abstract

The Belief-Desire-Intention (BDI) architecture is a popular framework for rational agents, yet most verification approaches are limited to analysing qualitative properties, for example whether an intention completes. BDI-based systems, however, operate in uncertain environments with dynamic behaviours: we may need quantitative analysis to establish properties such as the probability of eventually completing an intention. We define a probabilistic extension to the Conceptual Agent Notation (CAN) for BDI agents that supports probabilistic action outcomes, and probabilistic plan and intention selection. The semantics is executable via an encoding in Milner's bigraphs and the BigraphER tool. Quantitative analysis is conducted using PRISM. While the new semantics can be applied to any CAN program, we demonstrate the extension by comparing with standard plan and intention selection strategies (e.g. ordered or fixed schedules) and evaluating probabilistic action executions in a smart manufacturing scenario. The results show we can improve significantly the probability of intention completion, with appropriate probabilistic distribution. We also show the impact of probabilistic action outcomes can be marginal, even when the failure probabilities are large, due to the agent making smarter intention selection choices.

The BDI architecture, where agents are modelled based on their belief, desires, and intentions, provides a practical approach to developing intelligent agents. One of the key features of BDI agents is that they are able to pursue multiple intentions in parallel, i.e. in an interleaved manner. Most of the previous works have enabled BDI agents to avoid negative interactions between intentions to ensure the correct execution. However, to avoid execution inefficiencies, BDI agents should also capitalise on positive interactions between intentions. In this paper, we provide a theoretical framework where first-principles planning (FPP) is employed to manage the intention interleaving in an automated fashion. Our FPP approach not only guarantees the achievability of intentions, but also discovers and exploits potential common sub-intentions to reduce the overall cost of intention execution. Our results show that our approach is both theoretically sound and practically feasible. The effectiveness evaluation in a manufacturing scenario shows that our approach can significantly reduce the total number of actions by merging common sub-intentions, while still accomplishing all intentions.

The Belief-Desire-Intention (BDI) architecture is a popular framework for rational agents; most verification approaches are based on reasoning about implementations of BDI programming languages. We investigate an alternative approach based on reasoning about BDI agent semantics, through a model of the execution of an agent program. We employ Milner's bigraphs as the modelling framework and present an encoding for the Conceptual Agent Notation (CAN) language -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.

The BDI architecture, where agents are modelled based on their beliefs, desires, and intentions, provides a practical approach to developing intelligent agent systems. However, these systems either do not include any capability for first-principles planning (FPP), or they integrate FPP in a rigid and ad-hoc manner that does not define the semantical behaviour. In this paper, we propose a novel operational semantics for incorporating FPP as an intrinsic planning capability to achieve goals in BDI agent systems. To achieve this, we introduce a declarative goal intention to keep track of declarative goals used by FPP and develop a detailed specification of the appropriate operational behaviour when FPP is pursued, succeeded or failed, suspended, or resumed in the BDI agent systems. Furthermore, we prove that BDI agent systems and FPP are theoretically compatible for principled integration in both offline and online planning manner. The practical feasibility of this integration is demonstrated, and we show that the resulting agent framework combines the strengths of both BDI agent systems and FPP, thus substantially improving the performance of BDI agent systems when facing unforeseen situations.