PURPOSES : The purposes of this study are to identify appropriate numbers of drivers for different time periods by analyzing the service times of the Special Transportation System and to shorten the waiting time to within 15 minutes. METHODS : In this study, the service time is divided into the call connection time (At), dispatch time after reception (Bt), vehicle arrival time after dispatch (Ct), and vehicle boarding time (Dt), and the annual average value for each time zone is calculated by analyzing the dispatch system database. Furthermore, the number of drivers working in each time period is extracted and the appropriate number of drivers for ensuring the dispatch waiting time remains within 15 minutes is determined. RESULTS : It is more accurate to interpret the decrease in dispatches during lunchtime as a decrease in the number of operational vehicles owing to the drivers' lunchtimes rather than a decrease in demand. During lunchtime (as in previous studies) the number of operations decreases, but the average dispatch time (Bt) greatly increases to 22:42; thus, it cannot be seen as a decrease in dispatch demand. The number of operations during lunchtime is proportional to the number of drivers on duty. The number of drivers on duty is inversely proportional to the average dispatch time. If the number of drivers is increased by 11.6%, the average waiting time can be reduced to within 15 minutes. CONCLUSIONS : To resolve delayed call connection issues, we will introduce an artificial intelligence (AI) call center. During the hours of 7 PM to 6 AM, calls will mainly be handled by AI and the counseling personnel will switch to daytime work. We will also increase the number of drivers by 11.6% to ensure that the dispatch time does not exceed an average of 15 minutes after receiving a call. In particular, we will generate the work schedule such that more than 131 drivers work in the 12:00 to 13:00 hours during lunch time to improve the situation where users have to wait for a long time. To do this, we will overlap the work hours for 2 hours in Jeonju and 1 hour in other cities and counties. We have to increase the number of night shift workers from seven to 15 so that all cities and counties can operate vehicles 24 hours a day, 365 days a year.

PURPOSES : The objective of this study is to investigate the spatiotemporal factors influencing the waiting time of special transportation systems for disabled people in Seoul, Korea. METHODS : A parametric survival analysis was employed to identify the primary factors influencing longer waiting times, primarily because censoring data were included in the analysis. For the analysis, one-year historical data collected from the Seoul Call-taxi for the disabled in 2019 were used. Using normal probability plots and Anderson-Darling statistics, a log-normal model was estimated to fit the data. RESULTS : In terms of time, there was a time zone in which the service level in terms of waiting time, such as late-night and dawn, decreased depending on the time of day called by call taxi users with disabilities in Seoul. Moreover, the waiting time on weekends was shorter than on weekdays. Spatially, when a user requests a service, there are many nearby vehicles when the traffic conditions nearby are good, and the waiting time tends to decrease near the depot. CONCLUSIONS : The existing system needs to reflect the increase in vulnerable times and the improvement of the dispatch method, considering the difference in the primary purpose of travel on weekdays and weekends. Additionally, to resolve the phenomenon in which vehicles are concentrated in a specific area, waiting time can be improved using operational strategies that can uniformly distribute vehicles in space, such as spatial subdivision of depots, real-time monitoring, and relocation.