Planning with Perspectives -- Decomposing Epistemic Planning using Functional STRIPS

In this paper, we present a novel approach to epistemic planning called with perspectives (PWP) that is both more expressive and computationally efficient than existing state-of-the-art tools. Epistemic — knowledge belief essential in many multi-agent human-agent interaction domains. Most planners solve problems by either compiling propositional classical (for example, generating all possible atoms or formulae normal forms); explicitly encoding Kripke-based semantics. However, these methods become infeasible as problem sizes grow. decompose delegating reasoning about an external solver. We do modelling the using Functional STRIPS, which standard STRIPS supports use of external, black-box functions within action models. Building on recent work demonstrates relationship between what agent ‘sees’ it knows, define perspective each function, build solver for logic around this. Modellers can customise function agents, allowing new logics be defined without changing planner. ran evaluations well-known benchmarks compare planner, scenarios demonstrate expressiveness PWP approach. The results show our planner scales significantly better compared against, express succinctly.