Environmental policy in the context of complex systems: Statistical optimization and sensitivity analysis for ABMs

TL;DR

The paper tackles the computational bottleneck of using agent-based models (ABMs) for environmental policy design by coupling a statistical sensitivity analysis with ML-based policy optimization. It introduces a Gaussian process–based additivity test to assess whether the optimal policy x*(θ) depends on state variables θ, and a Bayesian optimization framework with Expected Improvement to efficiently search costly ABMs for θ-dependent optima. Applied to an extended Sugarscape ABM with pollution, the framework rapidly identifies interpretable policies that outperform baselines and provides dynamic analyses that connect with economic theory. This work demonstrates a practical path to harness ABMs for policy design and outlines directions for multi-objective optimization and empirical calibration.

Abstract

Coupled human-environment systems are increasingly being understood as complex adaptive systems (CAS), in which micro-level interactions between components lead to emergent behavior. Agent-based models (ABMs) hold great promise for environmental policy design by capturing such complex behavior, enabling a sophisticated understanding of potential interventions. One limitation, however, is that ABMs can be computationally costly to simulate, which hinders their use for policy optimization. To address this, we propose a new statistical framework that exploits machine learning techniques to accelerate policy optimization with costly ABMs. We first develop a statistical approach for sensitivity testing of the optimal policy, then leverage a reinforcement learning method for efficient policy optimization. We test this framework on the classic ``Sugarscape'' model, an ABM for resource harvesting. We show that our approach can quickly identify optimal and interpretable policies that improve upon baseline techniques, with insightful sensitivity and dynamic analyses that connect back to economic theory.

Figures (15)

• Figure 1: A basic framework for understanding the potential role of ABMs (and similar models) in policymaking. The arrows on the side of the figure emphasize that this is an iterative process in which policies are continually redesigned as new information about their effects is revealed and modeling techniques/computational power improves.
• Figure 2: Distribution of pollution, $t$ = 0
• Figure 3: Distribution of pollution, $t$ = 500
• Figure 4: Each boxplot is the distribution of agent survival rates across 100 model runs, of 500 steps each, of the sugarscape trade model.
• Figure 5: Each boxplot is the distribution of Gini coefficients (calculated from agent welfare) across 100 model runs, of 500 steps each, of the sugarscape trade model.