감마 분포 GLM을 활용한 고속도로 휴게소 매출액 추정 모형 개발

초록

Expressway service areas play a crucial role for road safety by addressing users' physiological needs and serving as evacuation sites during emergencies. Accurate revenue estimation is essential for developing appropriate and reasonable operational plans for these areas. However, predicting demand and revenue is challenging due to their unique characteristics: they are accessible only via expressways and are not the final destinations for users. Traditional quantitative methods, such as the gravity model, often fall short in addressing these challenges. Previous studies have proposed revenue estimation models, but they are difficult to apply during the planning stage because they rely on hard-to-obtain variables, such as entry rates and detailed traffic volumes. To address this limitation, this study introduces a new revenue estimation model for expressway service areas that uses easily accessible variables, making it suitable for use during the planning stages. We developed three Gamma-distribution generalized linear models using revenue data from 182 service areas from 2017 to 2021, which yielded Bayesian Information Criterion (BIC) values ranging from 8,001 to 8,226. After evaluating the model’s fit and verifying them with 2022 data, we identified the model with the best performance. Our analysis indicates that variables such as traffic volume, total floor area, express bus transfer service area, and distance to the previous service area positively impact revenue. Conversely, an excessively large gross floor area can negatively affect revenue. The findings from this study are expected to provide valuable insights for planning service areas in future expressway constructions.

키워드