基于混合蚁群算法的核应急车辆疏散路径规划  

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

作　　者：周怀芳 张华[1,2] 霍建文[1,2] 李林静[1,2] 陈波 林海涛 ZHOU Huaifang;ZHANG Hua;HUO Jianwen;Li Linjing;CHEN Bo;Lin Haitao(School of Information and Engineering,Southwest University of Science and Technology,Mianyang 621010,China;Sichuan Key Laboratory of Special Environmental Robotics,Southwest University of Science and Technology,Mianyang 621010,China)

机构地区：[1]西南科技大学信息工程学院,绵阳621010 [2]西南科技大学特殊环境机器人技术四川省重点实验室,绵阳621010

出　　处：《辐射研究与辐射工艺学报》2023年第6期65-78,共14页Journal of Radiation Research and Radiation Processing

基　　金：国家自然科学基金(12205245);四川省科技计划项目(2021YFG0376)资助。

摘　　要：核事故的发生具有不可预测性和破坏性,为应急车辆制定合理的疏散计划将危险区域的人员撤离至安置点,可以有效减少人员所受到的伤害。针对核事故下应急车辆路径规划问题,以累积辐射剂量为评价指标,提出了一种基于混合蚁群算法(Hybrid ant colony algorithm,HACO)的车辆路径规划方法。首先,利用模糊网络建立了时间窗内疏散路径平均通行时间期望模型,同时结合累积辐射剂量计算模型,建立了能够随时间变化的动态累积辐射剂量计算模型。然后在蚁群算法迭代过程中引入模拟退火算法,并且在邻域搜索中引入A*算法启发式思想,提高了算法全局寻优能力。为进一步提高算法的局部搜索能力,引入帕累托排序方式,在蚁群算法信息素更新方式中加入距离对信息素增量的影响。仿真结果表明:HACO算法相较于蚁群算法平均收敛值提高了31%,稳定性提高了30%,能够为核事故下疏散路径规划预案的制定提供技术支持。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.

关 键 词：核事故 路径规划 混合蚁群算法 动态模糊网络 累积辐射剂量模型 

分 类 号：TL72[核科学技术—辐射防护及环境保护] TL77