Development of estimated models of the number of potholes with the statistical optimization method

The objective of this paper is to determine a predictive model that uses the harmony search algorithm (HSA) based on available the multi-regression equation. The model employs the least squares method to predict the number of potholes in the Seoul metropolitan area. Independent variables were determined, based on traffic and weather data for each month in Seoul. Prior to the development of predictive models, empirical and stochastic factors that affect the occurrence of potholes were determined, resulting in a standardized regression coefficient from multi-linear regression analysis. A best-fit equation was derived from experiments using independent variables obtained from empirical and analytical approaches. The empirically and analytically filtered factors for each road management area in Seoul were used to develop the predictive models for the multiple regression analysis and the HSA. Fourteen predictive models were determined in this study. A performance comparison between these predictive models was conducted using the P-value, the root mean squared error, and the coefficient of determination.