Vehicle evacuation route planning in nuclear emergencies based on hybrid ant colony algorithm

Abstract

Nuclear accidents, although unpredictable and devastating, can be mitigated through well-formulated evacuation plans. An efficient evacuation of residents from hazardous zones to safer locations can be ensured through such plans. To address the vehicle path planning challenge under nuclear accidents, this paper proposes a method based on the hybrid ant colony algorithm (HACO). Cumulative radiation dose is used as a key assessment metric. Initially, a model estimating the average time for evacuating a route within a given time window is designed using a fuzzy network. In addition, a time-varying dynamic radiation dose model is proposed by incorporating the cumulative radiation dose calculation. The ant colony algorithm's iterative process is enhanced by the incorporation of the simulated annealing algorithm, while the heuristic approach of A* algorithm is employed for neighborhood searches. This integration results in an enhanced capacity for global optimization of the algorithm. For refining the local search capabilities of the algorithm, Pareto ordering is implemented. Additionally, the pheromone update method of the ACO algorithm is adjusted to account for the impact of distance on pheromone increments. Upon employing the HACO algorithm, simulation results indicate a 31% improvement in average convergence value and 30% boost in stability over the conventional ACO algorithm. These enhancements are instrumental in fortifying the planning of evacuation routes in the event of nuclear accidents.

Keywords

• nuclear accidents
• path planning
• hybrid ant colony algorithm
• dynamic fuzzy network
• cumulative radiation dose model