The increasing traveling demands are putting higher pressure on urban networks,
where the efficient driving modes highly depend on various non-intrusive ITS
equipment for interaction, which asks for higher maintenance scheduling plans
minimizing network loss. Current studies have researched methodologies with the
aspects of deterministic methods and metaheuristic algorithms under different
scenarios, but lack the simulation considering maintenance work type, urban
traffic characteristics as well as the ITS equipment. This study aims to
optimize the maintenance scheduling plan of urban ITS systems by using the
genetic algorithm (GA) and Dijkstra algorithm, as well as other judgmental
algorithms to minimize traffic delays caused by maintenance activities, and
presents a novel method to assess economic losses. A mixed integer programming
model is established simulating the real urban network while considering
multiple constraints, including the route selection principle, network updating,
network updating principle, etc. Then a complex urban network is randomly
assumed for the case study. Through case verification, the effectiveness of the
proposed model and algorithm in reducing the delay of the entire road network is
proved and reached a 19.5% loss avoidance compared to the traditional GA under
the case scenario. This study provides a theoretical basis and practical
guidance for the future maintenance and scheduling of intelligent transportation
systems in the environment of automatic driving, with advantages of
expandability, editability, and relatively high efficiency, but leaves
shortcomings of the possibility of falling into the local optima and traffic
assignment principle, which could be further studied in the algorithm and the
help with other advanced traffic assignment models.