[4P3-GS-10] AI application:

[4P3-GS-10-02] Interpretation through Balanced Tree Clustering Considering Interactions of Multiple Policies in System Dynamics

〇Yuta Sumiya¹, Naoyuki Echizenya¹, Tomoyuki Taguma¹, Kyosuke Toda¹, Takahisa Ando¹, Yoshiaki Oida¹, Yutaka Takahashi² (1. Fujitsu ltd., 2. Senshu University)

Keywords:System Dynamics, Monte Carlo Simulation, Clustering, Decision Tree

This study aims to provide a method for interpreting System Dynamics (SD) simulations using balanced tree clustering, focusing on analyzing nonlinear interactions and feedback loops in complex systems. While SD is a powerful tool for exploring dynamic behaviors, understanding variable impacts and identifying critical thresholds across numerous scenarios remains challenging. To address this, we show a new method to extract thresholds for policy variables by applying balanced tree clustering to the Earth4All model. The results reveal clear splits in clusters at specific policy values, offering a structured approach to interpreting SD scenarios. This method helps decision-makers identify critical factors and derive actionable guidelines for achieving desired outcomes.