SimDec Meets System Dynamics: Making Urban Planning Models Transparent

Urban planning is complex — and unpredictable.
Cities evolve through tangled interactions between people, housing, and jobs. Small policy shifts can trigger large, sometimes counterintuitive outcomes.

In “Extending System Dynamics Modeling Using Simulation Decomposition to Improve the Urban Planning Process” (Frontiers in Sustainable Cities, 2023), Julian Scott Yeomans and Mariia Kozlova show how Simulation Decomposition (SimDec) can bring clarity to that complexity — extending classical system dynamics models with visual sensitivity analysis that makes cause-and-effect relationships visible.

🔗 Read the full article: https://doi.org/10.3389/frsc.2023.1129316

The Challenge: Hidden Dynamics in Urban Models

Urban system models, like Forrester’s classic URBAN1, simulate how housing, population, and business activity evolve over time.
But these models often lack proper sensitivity analysis — the ability to tell which inputs (like immigration or construction rates) drive the results and how they interact.

Traditional Monte Carlo simulations capture uncertainty, but they produce only aggregate outcome distributions — without explaining why the outcomes occur.

That’s where SimDec changes everything.

How SimDec Extends System Dynamics

SimDec decomposes Monte Carlo outputs by linking every simulated outcome to the specific combination of input conditions that created it.
Instead of one colorless probability curve, we get a stacked histogram — where each layer represents a scenario, such as:

• High immigration and low housing availability,
• Low business growth and high population density,
• Or any other meaningful combination.

This visual decomposition transforms black-box models into transparent decision tools.

Case Study: The Urban Dynamics Model (URBAN1)

Yeomans and Kozlova applied SimDec to the URBAN1 model — a simplified but powerful representation of city evolution.

They varied three key inputs:

• Initial housing,
• Initial population, and
• Initial business structures.

Each was divided into “low” and “high” states, creating eight possible input scenarios.
The resulting visualization (Figure 1 in the paper) revealed striking contrasts:

• Cities starting with low housing grew faster — but were more prone to collapse.
• Higher starting values led to stability but slower expansion.

When the analysis focused on employment (labor force–to–jobs ratio), SimDec revealed that unemployment persisted under all scenarios, but its severity depended on how initial conditions combined.

Immigration, Outmigration, and Nonlinear Surprises

Next, the authors simulated uncertainty in immigration and outmigration rates.
The decomposed outputs (Figures 3–5) showed several nonlinear effects:

• Outmigration had a stronger influence on population than immigration,
• The system exhibited multiple “clusters” of outcomes — including total collapse in some scenarios,
• Feedback loops between housing and jobs shaped the ultimate stability of the city.

These patterns would remain invisible in standard time-series or single-scenario graphs.

Why It Matters for Urban Policy

SimDec equips planners and policymakers with a tool to:

• Visually explore how different policies affect urban resilience,
• Understand interactions among demographic, economic, and infrastructural factors,
• Communicate uncertainty intuitively to stakeholders.

In essence, it bridges quantitative modeling and qualitative understanding — turning complex simulation data into interpretable insight.

Conclusion

SimDec brings a new layer of interpretability to system dynamics.
It exposes how policies and uncertainties intertwine, helping urban decision-makers move from “What happens?” to “Why does it happen — and what can we do about it?”

Reference:
Yeomans, J. S., & Kozlova, M. (2023). Extending System Dynamics Modeling Using Simulation Decomposition to Improve the Urban Planning Process. Frontiers in Sustainable Cities, 5:1129316.
https://doi.org/10.3389/frsc.2023.1129316

‍