基于双层多粒子群算法的HDEHR系统资源分配

Resource allocation in HDEHR system based on bi-level multi-swarm PSO

为解决集中式区域医疗信息集成系统存在的访问瓶颈问题,设计了一种层次化分布式EHR集成模型HDEHR(hierarchical distributed EHR).介绍了HDEHR模型的总体架构,指出了该模型中的系统资源分配问题.为HDEHR系统资源分配问题进行了形式化定义,提出一种采用平均等待时间和网络与节点资源利用均衡度的调和模型作为最优化目标函数.根据HDEHR模型层次化的特点设计了一种双层多粒子群算法BLMSPSO(bi-level multi-swarm PSO),并用其解决了模型资源的最优化分配问题.采用HDEHR-Sim仿真平台进行了试验,对HDEHR算法和集中式EHR模型进行了比较.结果表明,该系统具备线性响应能力,能很好地解决集中式EHR集成模型存在访问瓶颈的问题.

To solve the bottleneck problem of centralized EHR integration model in regional health information integration, a hierarchical distributed EHR integration model （HDEHR） was proposed. The architecture of HDEHR was introduced to illuminate the resource allocation problem. The problem formulation was built and a reconcile model of minimum average waiting time and maximum network and node resource utilization was proposed as optimization objective function. According to the hierarchical characte- ristic of HDEHR model, the bi-level multi-swarm PSO algorithm （BLMSPSO） was designed to solve the optimization problem of resource allocation. Based on HDEHR-Sim platform, the simulation experiments were carried out to compare the performance of centralized EHR and HDEHR model. The results show that HDEHR has linear response capability, and can solve the bottleneck problem in centralized EHR model.