Forecasting Residential Gas Demand in an Aging Regional City Using a Hybrid Microsimulation Model

Japan’s super-aging society poses critical challenges for maintaining regional utility infrastructure, rendering conventional growth-oriented planning obsolete. Traditional top-down forecasting methods often fail to capture the heterogeneous impacts of changing household compositions and technological advancements on energy loads. This study addresses this gap by proposing a hybrid forecasting model that bridges demographic microsimulation with engineering-based efficiency modeling to predict residential gas demand in Okayama City through 2050. Utilizing daily gas consumption data from 892 households and Maximum a Posteriori (MAP) estimation, we identified baseline household attributes and simulated future demand by coupling demographic drivers with appliance stock turnover logic. The results reveal a non-linear demand trajectory characterized by a buffered contraction: an initial sharp decline driven by efficiency improvements (2020–2035), followed by a phase of stagnation (2035–2040) where the resilience of elderly households acts as a critical baseload buffer. The findings demonstrate that energy demand diverges non-linearly from total population trends, with single-elderly households emerging as the dominant load driver. This trajectory suggests that the transition toward carbon credit schemes and community-based service models is essential for converting this demographic stability into long-term economic viability. This study provides empirically grounded insights for rightsizing infrastructure and strategic business pivoting in shrinking regional cities.