大规模突发事件下医疗资源配置建模及算法

以医疗急救资源的配置问题为建模核心,运用次梯度最优算法对传统的拉格朗日松弛算法进行了改进。经迭代的拉格朗日乘数和改进LocAlloc算法对其上下限值的间距进行优化,加快了收敛速度,而贪婪算法可以连续完成对未覆盖的需求点的搜寻,实现了有效的医疗急救资源最优覆盖解。通过汶川地震的大规模医疗急救案例验证了该模型及算法的有效性和可行性。     

基于二维免疫的应急救援物资并行分配算法

现有应急救援物资分配研究大都局限于按照各发放点的优先级串行分配救援物资,难以适应复杂的大规模应急场景。为此,构建了面向多发放点的应急救援物资并行分配模型,并设计了基于二维二进制免疫和抗体修正的多发放点多救援物资并行分配算法。实验结果表明,该算法在大规模样本下较串行分配方法更加高效,能够从全局角度同时给出多个发放点的救援物资分配方案,为解决重大灾害中的复杂应急救援物资分配问题提供了一个有益的尝试。     

A novel hybrid method for improving ambulance dispatching response time through a simulation study

Response time is the most important factor in evaluating the performance of various Emergency Medical Services (EMS). In this paper, a novel hybrid method has been proposed to improve response time for ambulance dispatching. The proposed approach uses a linear hybrid metric based on network centrality measures, nearest neighbor method and first-in first-out (FIFO) policy. Other important parameters in ambulance dispatching such as the operating environment, rate of incoming emergency calls, available resources, hospitalization probability of the patients as well as distances and locations of units are all part of information used in this proposed approach. In line with the traditional metrics used in previous works, we have adopted a linear combined metric which is adjusted according to environment parameters. Results of extensive simulation experiments show reductions in response time by as much as 42% as compared to previous methods.     

Resilient control for wireless networked control systems under DoS attack via a hierarchical game

In this paper, the resilient control problem is investigated for a wireless networked control system (WNCS) under denial‐of‐service (DoS) attack via a hierarchical game approach. In the presence of a wireless network, a DoS attacker leads to extra packet dropout in the cyber layer of WNCS by launching interference power. A zero‐sum Markov game is exploited to model the interaction between the transmitter and the DoS attacker under dynamic network environment. Additionally, with the attack‐induced packet loss, an H_∞ minimax controller is designed in the physical layer by using a delta operator approach. Both value iteration and Q‐learning methods are used to solve the hierarchical game problem for the WNCS. The proposed method is applied to a load frequency control system to illustrate the effectiveness.