研究大模型树搜索规划中的成本意识，发现现有方法难找最优解
Cost-Awareness in Tree-Search LLM Planning: A Systematic Study
- 用统一框架对比四种树搜索算法，显式分析中间决策与失败模式
- 增加搜索计算不保证更优，双向搜索综合表现最佳，蒙特卡洛在短任务中最优
- 适合关注大模型规划效率与资源约束的科研人员和工程师
在资源受限条件下进行规划是现实决策的核心，但大多数大语言模型（LLM）规划器假设动作成本一致。我们系统地分析了树搜索型LLM规划器是否具备成本意识，以及能否高效生成预算可行的计划。相较于黑箱提示，显式的搜索树暴露了中间决策、节点评估和失败模式，使规划器行为可受控地进行消融分析。我们在统一框架下研究深度优先搜索、广度优先搜索、蒙特卡洛树搜索（MCTS）和双向搜索。实验表明，现有基于树的LLM规划器往往难以找到成本最优的计划，且额外的搜索计算并不能可靠提升最优性。在所评估的方法中，双向搜索在整体效率和成功率上表现最佳；MCTS在短时域任务中达到最高最优性。树搜索规划器因其推理步骤显式，特别适用于研究LLM规划机制，不同于通过后训练轨迹内化规划动态的普通LLM。研究结果表明，提升资源约束下的LLM规划能力可能需要新搜索算法，而非单纯扩大推理期计算资源。
原文摘要 · Abstract (English)
Planning under resource constraints is central to real-world decision making, yet most large language model (LLM) planners assume uniform action costs. We systematically analyze whether tree-search LLM planners are cost-aware and whether they efficiently generate budget-feasible plans. In contrast to black-box prompting, explicit search trees expose intermediate decisions, node evaluations, and failure modes, which allows for controlled ablations of planner behavior. We study depth-first search, breadth-first search, Monte Carlo Tree Search, and bidirectional search within a unified framework. Our experiments show that existing tree-based LLM planners often struggle to find cost-optimal plans, and that additional search computation does not reliably improve optimality. Among the methods evaluated, bidirectional search achieves the best overall efficiency and success rate. MCTS achieves the highest optimality on short-horizon tasks. Tree-search planners are especially valuable for studying LLM planning because their reasoning steps are explicit, in contrast to plain LLMs that internalize planning dynamics through post-training trajectories. Our findings suggest that improving LLM planning under resource constraints will likely require new search algorithms, rather than solely scaling inference-time compute.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。