Research on Simulation of Emergency Evacuation Outside Nuclear Power Plants Based on Genetic Algorithm and Variable Neighborhood Search Algorithm

Abstract This paper focuses on the evacuation process of the public during off-site emergency, and studies a simulation method for off-site emergency evacuation based on genetic algorithm and variable neighborhood search algorithm. And this model is used to calculate and analyze the off-site emergency evacuation plan for a nuclear power plant, based on information such as population, vehicles, assembly points, settle places, transportation networks, and radiation dose fields around the nuclear power plant. The optimization objectives are evacuation time, collective dose, or individual maximum effective dose. The consumption time of the evacuation process and the dose exposure received by personnel during the evacuation process are provided. And compare the results with traditional evacuation simulation algorithm. The results indicate that emergency evacuation plans should not only consider the evacuation time, but also comprehensively consider the dose received by personnel during the evacuation process. In addition, this paper analyzed the impact of adding private cars on the evacuation process, and the results showed that when adding more than 50% of private cars to the number of buses, the impact on evacuation time, collective dose, and individual maximum effective dose significantly increased. In addition, when the evacuation time is taken as the optimization objective, the evacuation time of the optimized evacuation plan through genetic algorithm combined with variable neighborhood search algorithm is shorter than that of the traditional simulated evacuation plan. This algorithm can generate a more optimized evacuation plan.