For the problem of global path optimization in large-scale public indoor scenarios involving a robot navigating among multiple targets, a rapid response optimization algorithm was studied to address the criteria of minimizing energy consumption, travel time, and distance. This algorithm combines the strong local search capability of the Union-Variable Neighborhood Descent (Union-VND) algorithm with the good global convergence of the Ant Colony Optimization (ACO) algorithm. To mitigate the conflict between population diversity and convergence speed in ACO, modifications were made to the state transition equation, information pheromone update, introducing pseudo-random state transitions and distance-dependent reward-penalty system. Additionally, five neighborhood search methods from Union-VND were integrated into the improved ACO, ensuring global optimality, accuracy, and faster convergence. When applied to multi-target path planning for a manned robot in large indoor public environments, the combined algorithm demonstrated superior optimization of the stated criteria, fastest convergence, and rapid response in determining the optimal sequence of destinations, showcasing practical value.