Dynamic bus dispatching under the environment of Internet of things

Public transit is one of the most important service sectors of a modern city to satisfy people's intra-city transportation requirements. Dynamic bus dispatching in transit system plays a very important role since it is directly related to the daily operations of a transit agency. The main difficulty of dynamic bus dispatching lies in the fact that a lot of exceptional events or factors are required to be efficiently processed in transit system based on the current bus running plan and crew allocation plan. In the literature papers on dynamic bus dispatching, the assumption that bus timetables on stops are pre-determined may not be true for many developing countries, and a set of state space variables corresponding to real time environment information related to transit system is not fully integrated into the dynamic bus dispatching models. In this paper, a new approach for dynamic bus dispatching under the Internet of things environment is proposed. Based on the real time passenger flows and bus arriving time on stops, an optimization model with the objective of minimizing total passenger waiting time is established. A memory-based genetic algorithm is developed to solve the model under dynamically evolved environment. The effectiveness of the proposed approach is verified under different scenarios via simulation study.