# 通过生成式轨迹规划实现基于模型的在线决策

> 原文链接：https://robotics.hellobike.com/papers/2026-10-13-2610-18677
> 频道：papers | 更新日期：2026-10-13 | 标签：arxiv, cs.RO
> 核心摘要：GTP用扩散先验生成并结合学习的动力学、奖励和值模型细化动作序列，在连续控制和长时程操作任务上验证在线决策。

# 通过生成式轨迹规划实现基于模型的在线决策

::: v-pre
> 原文链接：[Model-Based Online Decision Making via Generative Trajectory Planning](https://arxiv.org/abs/2610.18677)（arXiv 2610.18677 · 发布日期：2026-10-13）

**Model-Based Online Decision Making via Generative Trajectory Planning**

GTP用扩散先验生成并结合学习的动力学、奖励和值模型细化动作序列，在连续控制和长时程操作任务上验证在线决策。

## 核心要点

- 在线决策需兼顾高效推理与长时程推理，现有价值法和轨迹优化各有限制。
- GTP学习动作序列的生成模型作为提议分布，避免每步从头优化序列。
- 候选序列由扩散先验采样，并用学习的动力学、奖励和值模型细化。
- 在连续控制和长时程操作任务中，性能匹配或超过多类先进基线，并在基线退化时保持稳定。

**验证范围与局限**：验证范围为连续控制和长时程操作任务；摘要未说明仿真或真机设置。

## 英文摘要 (Abstract)

Online decision-making faces a tension between efficient inference and long-horizon reasoning. Value-based methods rely on short-horizon bootstrapping that limits the propagation of long-term information, while trajectory optimization methods plan explicitly but must solve a high-dimensional optimization problem at every decision step. We propose Generative Trajectory Planning (GTP), a model-based framework that addresses this tension by learning a generative model as a proposal distribution over action sequences. Instead of optimizing sequences from scratch, GTP samples candidate sequences from a diffusion prior and refines them using learned dynamics, reward, and value models, enabling efficient trajectory-level reasoning. Across continuous control and long-horizon manipulation tasks, GTP matches or exceeds state-of-the-art model-free, diffusion-policy, and model-based baselines, and remains stable in regimes where strong baselines degrade. Our results suggest that generative models are best understood not as policies, but as structured proposal distributions for planning, offering a new perspective on how to integrate generative modeling with model-based decision-making.

作者：Haldun Balim, Yilun Du, Na Li
:::

## 原始参考来源 (Sources)

- [arXiv 2610.18677：Model-Based Online Decision Making via Generative Trajectory Planning](https://arxiv.org/abs/2610.18677)