Bus Driver Scheduling System via Genetic Algorithm with Enhanced Crossover Mechanism

The thesis investigates the problem and constraints related to bus driver scheduling in a case study. The unfairness of job distribution and the ineffectiveness of break-time assignment among bus drivers are factors identified as problem issues in the bus driver management. Hence, the objective of the study is to develop a model to solve the bus driver scheduling problem (BDSP). Among the approaches reviewed for solving this BDSP, genetic algorithm (GA) has been identified as the most potential solution approach. In the proposed GA, horizontal crossover with multiple-point and directed mutation techniques are introduced in its natural representation as part of the approach. The solutions obtained show that the proposed GA technique is able to improve the quality of solutions. The model has efficiently solved the bus driver scheduling problem in the case of Universiti Utara Malaysia (UUM). The proposed solution approach is able to generate quality schedule
efficiently and quickly when compared to the human-generated schedule.