Resumen
Based on a detailed GPS-Data Analysis for a free-floating Bike Sharing System, mobility patterns of the bike usage were identified spatially and temporally. Depending on different factors like seasons, weather conditions, time of the day and holidays/weekends, we built up a demand model in order to forecast the upcoming demand at certain time and place. This model reveals optimal fleet distributions for different zones in the operating area and for different time slots. To redistribute the fleet a reasonable relocation strategy was created in order to obtain an optimal distribution of bikes within the operating area. Two different kinds of relocation methods are evaluated: an operator-based relocation strategy, where at least one vehicle redistributes a part of the fleet. Additionally, a user-based relocation scheme was created: based on pricing benefits for the user, at least a share of the needed relocations are conducted by users, almost cost-neutral. There is a certain threshold though, when the necessary fleet relocations are too urgent and an operator-based intervention is inevitable. At the end, we carried out a validation to show what kind of effects a real life relocation could imply.