---
id: 20260904-T0-14
title: "BCO优化：让Agent学会用世界模型自我改进，超越传统优化器"
title_en: "BCO: Explicit World Model for Agent Optimization"
url: https://ai.daily.yangsir.net/daily/20260904-T0-14
issue_date: 2026-09-04
publish_date: 2026-09-03T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2609.01861
---

# BCO优化：让Agent学会用世界模型自我改进，超越传统优化器

一篇新论文提出信念校准优化（BCO），这是一种为LLM智能体设计的显式世界模型优化框架。该方法让智能体不仅能评估当前分数，还能模拟和预测不同修改策略的未来影响，从而更有效地迭代自己的源代码。实验表明，该框架优于传统的基于编码器Agent的优化方法。

## English Version

**BCO: Explicit World Model for Agent Optimization**

A new paper presents Belief-Calibrated Optimization (BCO), an explicit world model framework for optimizing LLM agents. This allows agents to not only evaluate current scores but also simulate and predict the future impact of different modification strategies, leading to more effective self-iteration. Experiments show the framework outperforms traditional coding-agent optimizers.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2609.01861)

**详情页**：https://ai.daily.yangsir.net/daily/20260904-T0-14

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*