To address the traffic efficiency and safety issues arising from mandatory lane-changing in lane reduction sections of urban roads within the vehicle-infrastructure collaboration environment, an automated vehicle cooperative lane-changing control strategy is proposed based on game theory concepts. A two-player non-cooperative game model with complete information is developed for a multi-lane reduction scenario. This model is expanded to include a transformation of cooperative-non-cooperative games, based on the overall reasonableness condition and individual reasonableness condition judgment. Passenger driving preferences are introduced into the safety and efficiency gain functions to accommodate the driving style inclinations of various passengers and enhance the human-like decision-making of vehicle lane changes. Subsequently, the concept of gain equivalence is used to resolve the Nash equilibrium of the game system and allocate revenue based on Shapley value to determine the optimal control strategy for the game vehicles in the two models. Finally, the joint simulation method of SUMO/TraCI and Python is employed to experimentally validate typical urban lane reduction scenarios. The experimental results demonstrate that the cooperative lane-changing control strategy, based on two types of game models, is effective in moderate to high traffic demand levels and exhibits significant local effects in the lane reduction zone. This control strategy not only effectively enhances traffic safety and efficiency in the lane reduction road zone but also has the capability to suppress the formation and spread of traffic shock waves, thereby increasing the stability of traffic flow. It can serve as a valuable reference for the mandatory lane-changing control of automated vehicles in the future vehicle-infrastructure collaboration environment.