City Editing: Hierarchical Agentic Execution for Dependency-Aware Urban Geospatial Modification

TL;DR

This work forms urban renewal as a machine-executable task that iteratively modifies existing urban plans represented in structured geospatial formats and introduces an iterative execution-validation mechanism that mitigates error accumulation and enforces global spatial consistency during multi-step editing.

Abstract

As cities evolve over time, challenges such as traffic congestion and functional imbalance increasingly necessitate urban renewal through efficient modification of existing plans, rather than complete re-planning. In practice, even minor urban changes require substantial manual effort to redraw geospatial layouts, slowing the iterative planning and decision-making procedure. Motivated by recent advances in agentic systems and multimodal reasoning, we formulate urban renewal as a machine-executable task that iteratively modifies existing urban plans represented in structured geospatial formats. More specifically, we represent urban layouts using GeoJSON and decompose natural-language editing instructions into hierarchical geometric intents spanning polygon-, line-, and point-level operations. To coordinate interdependent edits across spatial elements and abstraction levels, we propose a hierarchical agentic framework that jointly performs multi-level planning and execution with explicit propagation of intermediate spatial constraints. We further introduce an iterative execution-validation mechanism that mitigates error accumulation and enforces global spatial consistency during multi-step editing. Extensive experiments across diverse urban editing scenarios demonstrate significant improvements in efficiency, robustness, correctness, and spatial validity over existing baselines.

Figures (9)

• Figure 1: Traditional Urban Planning vs. Agentic City Editing. Top: Traditional urban planning generates urban layouts from scratch. Bottom: City editing incrementally refines existing urban environments in response to natural language-based human instructions while preserving structural constraints.
• Figure 2: Overview of CEAE. Given an existing urban plan and a natural-language editing instruction, the framework decomposes the editing process into structured intents and executes them in a staged, geometry-aware manner, where intermediate results are iteratively validated and propagated across stages to ensure reliable incremental refinement and produce an updated urban geospatial layout.
• Figure 3: Ablation study on the impact of task planner agent ($-t$) and validator agent ($-v$) on execution accuracy and robustness.
• Figure 4: Robustness evaluation results across three task levels, measured by 1-REE / 1-ACE and EVR.
• Figure 5: Hyperparameter sensitivity of the maximum re-execution attempts $R$.