Optimisation of new energy vehicle traffic flow and application of hybrid multi-objective evolutionary algorithm based on internet of things simulation of urban mobility simulation platform

The current optimisation of traffic flow faces challenges, such as increased congestion at intersections and the inability of traditional signal control strategies to adapt to dynamic mixed traffic. With the increasing proportion of new energy vehicles, the exhaust emission model needs to be reconstructed, and existing methods often focus on single objective optimisation, ignoring the synergistic trade-off between delay and emissions. This paper selects a multi-objective optimisation (MOO) model consisting of average vehicle delay and average exhaust emissions. It uses backpropagation neural network (BPNN) to improve non dominated sorting genetic algorithm II (NSGA-II), and utilises urban traffic simulation (SUMO) traffic model to establish and optimise micro road traffic models of intersections. According to the analysis of the optimisation results, the signal cycle at the optimised intersection has been shortened by 16.4%, the average delay time has been reduced by 14.9%, and the exhaust emissions have been reduced by 8.5%.