GenePlan: Evolving Better Generalized PDDL Plans using Large Language Models

2026-03-10

AI Summary

This work proposes a novel approach to automatically synthesize efficient and generalizable domain-specific planners for classical PDDL planning tasks. By formulating generalized planning as an optimization problem, the method integrates large language models (e.g., GPT-4o) with evolutionary algorithms to iteratively evolve interpretable Python programs that serve as planners. Evaluated on eight standard domains, the resulting planners achieve an average SAT score of 0.91—approaching the performance of state-of-the-art planners (0.93) and substantially outperforming LLM-based baselines such as chain-of-thought prompting (0.64). Moreover, the synthesized planners solve new instances in just 0.49 seconds on average, with a generation cost of approximately $1.82 per domain, demonstrating superior trade-offs among performance, speed, and cost compared to existing LLM-driven methods.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

We present GenePlan (GENeralized Evolutionary Planner), a novel framework that leverages large language model (LLM) assisted evolutionary algorithms to generate domain-dependent generalized planners for classical planning tasks described in PDDL. By casting generalized planning as an optimization problem, GenePlan iteratively evolves interpretable Python planners that minimize plan length across diverse problem instances. In empirical evaluation across six existing benchmark domains and two new domains, GenePlan achieved an average SAT score of 0.91, closely matching the performance of the state-of-the-art planners (SAT score 0.93), and significantly outperforming other LLM-based baselines such as chain-of-thought (CoT) prompting (average SAT score 0.64). The generated planners solve new instances rapidly (average 0.49 seconds per task) and at low cost (average $1.82 per domain using GPT-4o).

Problem

• generalized planning
• PDDL
• domain-dependent planner
• plan generalization
• classical planning

Innovation

• Generalized Planning
• Large Language Models
• Evolutionary Algorithms
• PDDL
• Interpretable Planners