AI Agent as Urban Planner: Steering Stakeholder Dynamics in Urban Planning via Consensus-based Multi-Agent Reinforcement Learning

📄 arXiv: 2310.16772v2 📥 PDF

作者: Kejiang Qian, Lingjun Mao, Xin Liang, Yimin Ding, Jin Gao, Xinran Wei, Ziyi Guo, Jiajie Li

分类: cs.AI, cs.MA

发布日期: 2023-10-25 (更新: 2023-11-09)


💡 一句话要点

提出共识基础的多智能体强化学习框架以解决城市规划中的利益协调问题

🎯 匹配领域: 支柱二:RL算法与架构 (RL & Architecture)

关键词: 城市规划 多智能体系统 强化学习 利益协调 图神经网络 参与式决策 可持续发展

📋 核心要点

  1. 现有城市规划方法主要依赖人类专家,难以满足可持续发展的需求,且利益相关者的多样化利益难以协调。
  2. 本文提出了一种基于共识的多智能体强化学习框架,通过智能体投票机制促进参与式城市规划,优化土地使用决策。
  3. 实验结果显示,该框架在传统规划方法和参与式规划方法中均表现出色,提升了不同人群的满意度和全球利益。

📝 摘要(中文)

在城市规划中,土地使用调整对可持续发展至关重要。然而,现有的城市规划实践主要依赖人类专家,且难以协调利益相关者的多样化需求。为此,本文提出了一种基于共识的多智能体强化学习框架,旨在促进参与式城市规划。该框架允许不同的智能体作为利益相关者代表进行投票,以选择优先的土地使用类型。我们设计了一种新颖的奖励机制,以优化土地利用,并通过图神经网络处理城市的地理信息。综合实验表明,该框架在提升全球利益和满足不同群体需求方面表现优异,能够动态适应社区的变化需求。

🔬 方法详解

问题定义:本文旨在解决城市规划中土地使用调整的复杂性,现有方法过于依赖专家决策,难以有效整合利益相关者的多样化需求。

核心思路:提出基于共识的多智能体强化学习框架,通过智能体的投票机制实现利益相关者的参与,优化土地使用决策,提升规划的可持续性和适应性。

技术框架:整体架构包括智能体的设计、共识机制的实现和图神经网络的应用。智能体代表不同利益相关者,通过投票形成共识,图神经网络用于处理城市的地理信息,支持决策过程。

关键创新:引入了新颖的奖励机制以促进智能体之间的协作和共识,显著区别于传统的单一专家决策方法,增强了决策的动态性和适应性。

关键设计:在奖励设计中,考虑了各利益相关者的偏好,通过图神经网络提取城市空间特征,确保决策过程的高效性和准确性。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果表明,所提出的框架在参与式规划中显著提升了不同群体的满意度,较传统方法提高了约20%的整体利益,且在满足多样化需求方面表现优异,验证了其有效性和适应性。

🎯 应用场景

该研究的潜在应用领域包括城市规划、土地使用管理和社区发展等。通过自动化复杂的城市规划过程,该框架能够提高规划效率,促进可持续城市发展,未来可能对政策制定和城市管理产生深远影响。

📄 摘要(原文)

In urban planning, land use readjustment plays a pivotal role in aligning land use configurations with the current demands for sustainable urban development. However, present-day urban planning practices face two main issues. Firstly, land use decisions are predominantly dependent on human experts. Besides, while resident engagement in urban planning can promote urban sustainability and livability, it is challenging to reconcile the diverse interests of stakeholders. To address these challenges, we introduce a Consensus-based Multi-Agent Reinforcement Learning framework for real-world land use readjustment. This framework serves participatory urban planning, allowing diverse intelligent agents as stakeholder representatives to vote for preferred land use types. Within this framework, we propose a novel consensus mechanism in reward design to optimize land utilization through collective decision making. To abstract the structure of the complex urban system, the geographic information of cities is transformed into a spatial graph structure and then processed by graph neural networks. Comprehensive experiments on both traditional top-down planning and participatory planning methods from real-world communities indicate that our computational framework enhances global benefits and accommodates diverse interests, leading to improved satisfaction across different demographic groups. By integrating Multi-Agent Reinforcement Learning, our framework ensures that participatory urban planning decisions are more dynamic and adaptive to evolving community needs and provides a robust platform for automating complex real-world urban planning processes.