Journal of Korean Society of Transportation ISSN:1229-1366(Print) 2234-4217(Online) 대한교통학회지

The importance of non-motorized transportation has increased in urban transportation systems as interest in environmental issues such as climate change has arisen. The public bike-sharing system in Seoul was born as part of efforts to improve the issues. Another problem, however, has occurred in the use of bikes due to the imbalance of demand-supply of public bikes at each station. In this study, an optimal rebalancing strategy was established based on the estimated rent-return demand of each station in the purpose of keeping the inventory at a certain level. The optimized model by MATLAB programming uses the integer programming method to estimate the number of bike. The strategy consists of three stages. First, drop-off stations were selected through estimated rent-return demand. Second, clustering method was implemented to improve the efficiency of the process of finding the shortest rebalancing route. Third, the direction of the route is determined centering around the drop-off stations, and optimized route is proposed by selecting nearest pick-up stations for the purpose of fulfilling drop-off. The study was conducted using the data of bike rent-return history in Mapo-gu, Seoul. As a result, the demand satisfaction ratio was 0.82 with 112 of the 136 data were satisfied on the 2-hour peak time. The propriety outcome of the number of docks was 0.47 in that 32 docks out of 68 are appropriately allocated. The coefficient of variation was improved from 1.38 to 1.12 as an index of uniformity of inventory margin. The standard deviation of the frequency distribution curve of inventory margin decreased from 24.9 to 15.7, and F-test was implemented to confirm significance.