SimCity: Multi-Agent Urban Development Simulation with Rich Interactions

TL;DR

SimCity tackles the problem of modeling macroeconomic dynamics with heterogeneous agents in an urban context by leveraging LLM-driven households, firms, a central bank, and a government within a mapped city. The core method combines an agent-based, macroeconomic framework with visualized geography and a vision-language model to place firms, enabling analysis of canonical phenomena like the Phillips Curve, Okun's Law, and the Beveridge Curve, along with urban expansion dynamics. Key contributions include the LLM-based multi-agent architecture, a spatially rendered city environment, and a systematic evaluation checklist showing SimCity reproduces established stylized facts while remaining robust to stochastic variation. The work advances practical insights for interpreting macroeconomic regularities in spatially embedded, adaptive economies and lays groundwork for more realistic, interactive urban simulations.

Abstract

Large Language Models (LLMs) open new possibilities for constructing realistic and interpretable macroeconomic simulations. We present SimCity, a multi-agent framework that leverages LLMs to model an interpretable macroeconomic system with heterogeneous agents and rich interactions. Unlike classical equilibrium models that limit heterogeneity for tractability, or traditional agent-based models (ABMs) that rely on hand-crafted decision rules, SimCity enables flexible, adaptive behavior with transparent natural-language reasoning. Within SimCity, four core agent types (households, firms, a central bank, and a government) deliberate and participate in a frictional labor market, a heterogeneous goods market, and a financial market. Furthermore, a Vision-Language Model (VLM) determines the geographic placement of new firms and renders a mapped virtual city, allowing us to study both macroeconomic regularities and urban expansion dynamics within a unified environment. To evaluate the framework, we compile a checklist of canonical macroeconomic phenomena, including price elasticity of demand, Engel's Law, Okun's Law, the Phillips Curve, and the Beveridge Curve, and show that SimCity naturally reproduces these empirical patterns while remaining robust across simulation runs.

Figures (8)

• Figure 1: The framework of SimCity. Left: A visualized map with three types of buildings. Right: The rich interactions between various agent modules.
• Figure 2: Emergence of the Phillips Curve and Okun’s Law in SimCity simulations. $r$-value is the Pearson correlation coefficient and $p$-value indicates the statistical significance of it.
• Figure 3: Beveridge Curve and other macroeconomic emergences from SimCity.
• Figure 4: The results from different random seeds demonstrate that the observed regularity is robust.
• Figure 5: GDP & population curves, and map changes during the move-in phase.