Behaviour Planning: A Toolkit for Diverse Planning

2024-05-07

arXiv.org

AI Summary

Existing diverse planning approaches rely heavily on distance-function-based modeling, suffering from poor interpretability and limited expressive power. To address this, we propose a behavior-oriented planning framework that introduces— for the first time—a modular diversity modeling architecture and a qualitative diversity description language, enabling semantic modeling and attributional explanation of behavioral differences among plans. Methodologically, we integrate formal diversity modeling with Satisfiability Modulo Theories (SMT) constraint solving to precisely encode and efficiently satisfy diversity requirements. Experimental results demonstrate that, compared to state-of-the-art methods, our framework achieves significant improvements in diversity quality, solution feasibility, and computational efficiency. It further offers strong interpretability, high flexibility, and cross-domain generality, effectively supporting real-world applications such as plan recognition and business process automation.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Diverse planning is the problem of generating plans with distinct characteristics. This is valuable for many real-world scenarios, including applications related to plan recognition and business process automation. In this work, we introduce emph{Behaviour Planning}, a diverse planning toolkit that can characterise and generate diverse plans based on modular diversity models. We present a qualitative framework for describing diversity models, a planning approach for generating plans aligned with any given diversity model, and provide a practical implementation of an SMT-based behaviour planner. We showcase how the qualitative approach offered by Behaviour Planning allows it to overcome various challenges faced by previous approaches. Finally, the experimental evaluation shows the effectiveness of Behaviour Planning in generating diverse plans compared to state-of-the-art approaches.

Problem

• Expresses diversity beyond distance functions in planning
• Enables customizable n-dimensional grid-based diversity modeling
• Supports non-classical planning categories like over-subscription

Innovation

• Uses n-dimensional grid for expressive diversity modeling
• Introduces customizable Behavior Planning toolkit
• Supports planning categories beyond classical planning