Bi-objective bus scheduling optimization with passenger perception in mind.

With the development of big traffic data, bus schedules should be changed from the traditional "empirical" rough scheduling to "responsive" accurate scheduling to meet the travel needs of passengers. Based on passenger flow distribution, considering passengers' feelings of congestion and waiting time at the station, we establish a Dual-Cost Bus Scheduling Optimization Model (Dual-CBSOM) with the optimization objectives of minimizing bus operation and passenger travel costs. Improving the classical Genetic Algorithm (GA) by adaptively determining the crossover probability and mutation probability of the algorithm. We use an Adaptive Double Probability Genetic Algorithm (A_DPGA) to solve the Dual-CBSOM. Taking Qingdao city as an example for optimization, the constructed A_DPGA is compared with the classical GA and Adaptive Genetic Algorithm (AGA). By solving the arithmetic example, we get the optimal solution that can reduce the overall objective function value by 2.3%, improve the bus operation cost by 4.0%, and reduce the passenger travel cost by 6.3%. The conclusions show that the Dual_CBSOM built can better meet the passenger travel demand, improve passenger travel satisfaction, and reduce the passenger travel cost and waiting for cost. It is demonstrated that the A_DPGA built in this research has faster convergence and better optimization results.