Flexible crowd evacuation optimization in public places: A cell-based approach considering adjustable facilities

Crowds are becoming increasingly dense in large public places, leading to congestion especially during peak periods. It is crucial to propose a crowd evacuation optimization strategy to alleviate the congestion. Considering the impact of advanced facilities on pedestrian travel, this paper incorporates the flow dynamics of the cell transmission model and the adjustability of facilities. An integrated pedestrian flow assignment and facility configuration adjustment model is established for a crowd evacuation strategy under the system optimal criterion. To efficiently solve the proposed mixed-integer linear programming model, an improved Benders decomposition algorithm with a fixing strategy is developed. In this algorithm, the subproblem is reconstructed based on a time-expanded network, and a branch-and-cut approach is implemented to enhance the master problem. Additionally, to analyze the gap between system optimum and user equilibrium pedestrian traffic conditions, a potential-based pedestrian flow loading algorithm is introduced to obtain a stochastic dynamic user equilibrium flow pattern given the optimized facility configuration. According to the numerical example of a metro station, the computational time and the gap are computed to validate the efficiency of the algorithm under various experimental settings. The numerical results show that, compared to the user equilibrium criterion, the flow pattern and facility configuration under the system optimal criterion can decrease the total travel time by 4.1%-10.6% and alleviate congestion at critical bottlenecks by 31.1%. The research findings are ultimately summarized as an enhanced guidance map, demonstrating the potential application of our approach in developing an intelligent crowd evacuation system for smart cities.