The Belief-Desire-Intention Ontology Design Pattern

Belief-Desire-Intention (BDI) is a widely recognised framework for modelling rational agents' mental states in artificial intelligence and cognitive sciences. However, the integration of BDI concepts into structured knowledge representations remains challenging, limiting interoperability and reuse across domains. This paper introduces a modular Ontology Design Pattern (ODP) for modelling BDI mental states, ensuring semantic precision and adaptability. The proposed ODP captures the fundamental components of BDI—beliefs, desires, and intentions—along with their interrelations, dynamic evolution, and contextual dependencies. By leveraging existing ontologies and aligning with best practices in ontology design, the pattern enables the representation of agent reasoning, decision-making processes, and goal-oriented behaviours. We demonstrate the utility of the BDI ODP through different application scenario, such as autonomous systems and tasks that LLMs base on BDI, showcasing its potential to enhance reasoning, communication, and system interoperability in neuro-symbolic systems. This work contributes to the ontology engineering field by providing a reusable and extendable model for integrating mental state reasoning into diverse knowledge representation systems.