Research on Public Vehicle Evacuation Path Planning Model Based on Spatiotemporal Network

Abstract

This study aims to improve the efficiency of public vehicle evacuation during large-scale disasters by minimizing travel and waiting times for individuals and vehicles. To accomplish this, an S-curve behavior model was used to estimate evacuation demand, and a network model was developed to consider temporal and spatial factors of gathering points. A hybrid genetic algorithm and simulated annealing approach were utilized with an "enumerate then optimize" strategy and a step to temporarily retain optimal solutions for refinement. The effectiveness of the proposed model and algorithms was demonstrated in a case study of a typhoon evacuation in Chikan District, providing valuable insights for urban evacuation planning.