In recent years, numerical simulation approaches have gained increasing attention for evaluating policies and strategies in local governments and private organizations. These methods can be valuable for capturing the dynamic interactions and assessing policy impacts under different scenarios. However, practical use of these methods continues to face challenges, particularly in defining appropriate system boundaries that accurately capture the scope and complexity of real-world problems. This study proposes a framework called Model-based Model Design (MBMD) to support the formulation and validation of system boundaries in simulation-based policy evaluation. MBMD enables a structured assessment of simulation models by systematically describing the target system's components, environment, functions, and performance metrics, and by analyzing the validity of use cases derived from this structure. A simple case study involving municipal policy simulations demonstrates the practical applicability of the approach and highlights how MBMD supports transparent and rational boundary-setting, which is essential for credible policy evaluation.
Nakashima et al. (Wed,) studied this question.