An improved optimization model for crowd evacuation considering individual exit choice preference

Fei Gao,Zhiqiang Du,Martin Werner,Yuhui Zhao
DOI: https://doi.org/10.1111/tgis.12984
IF: 2.568
2022-09-06
Transactions in GIS
Abstract:Guidance‐assisted crowd evacuation is a process of combining individual exit choice behavior with managers' exit assignment control. The knowledge of individual exit choice preference is of great significance for optimizing global exit assignment planning. This study proposes an improved optimization model for crowd evacuation by integrating the individual‐level exit choice preference analysis with system‐level exit assignment optimization to represent more realistic crowd evacuation decisions. First, the impact factors of individual exit choice behavior are considered in a mixed logit model to predict the probability of each individual choosing each exit in specific situations. Second, a preference‐based exit filtering strategy is designed to analyze the sensible alternative exits for individuals or groups in multi‐scale evacuation cells. Finally, to pursue optimal exit assignment planning, a multi‐objective particle swarm optimization algorithm and an improved social force model are adopted to simulate the process of crowd evacuation and evaluate the performance of the specific exit assignment plans. The case study of an outdoor multiple‐exit scenario in Xi'an, China, indicates that the proposed model can help managers to understand the heterogeneity of individual evacuation behaviors. Furthermore, it will support more reliable and realistic evacuation decisions in real‐life situations than conventional plans that typically implement the top‐ n strategy.
geography
What problem does this paper attempt to address?