A constrained self-adaptive harmony search algorithm (CSAHS) with 2-opt swapping (CSAHS-2opt) for driver scheduling problem of university shuttle bus (DSPUSB) was proposed in this study. In generating a schedule, there are some constraints that should be catered to obtain an optimum result. In DSPUSB, fairness of task distribution among drivers is one of the main issues that are very crucial and can affect the performance and quality of services. Therefore, to maximise the fairness, the violation of soft constraints for DSPUSB including shift and route should be minimised. CSAHS with 2-opt swapping was proposed focusing on minimising soft constraint violation. In standard harmony search (HS), the value of distance bandwidth (BW) parameter was static, while in this study the BW was dynamically changed and determined based on the current solution of each driver every week. Here, a set of distance BW value was formed based on shift constraint. In each iteration, the BW values were often changed and randomly chosen within the set, whereas the 2-opt swapping normally used in travelling salesman problem was applied for route constraint based on some rules. These improvements are capable of reducing the repetition task that leads to fairness issue. The result demonstrated that CSAHS-2opt gave better solutions compared with standard HS, improved HS and parameter-adaptive HS. © 2018, King Fahd University of Petroleum & Minerals.