军事信息系统服务资源分配并行优化方法  

Parallel optimization method of service resource allocation in military information system

作　　者：于晓浩[1,2] 胡丹[3] 罗雪山[2] 刘俊先[2] 

机构地区：[1]北京系统工程研究所,北京100101 [2]国防科技大学信息系统工程重点实验室,长沙410073 [3]空军工程大学电讯工程学院,西安710077

出　　处：《系统工程理论与实践》2012年第9期2078-2086,共9页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(70601036);"十一五"装备预先研究项目

摘　　要：针对如何提高面向服务军事信息系统中任务工作流执行的时效性和成功概率,提出了服务资源分配的并行优化方法.首先给出了服务资源分配的系统框架、在分析服务并行执行数目、任务成功率、任务完成时间及服务执行代价之间关系的基础上,建立了服务并行优化的目标规划数学模型,并提出了一种求解该模型的改进粒子群算法(DPSO).该算法通过引入粒子细微扰动、优化粒子飞行边界及粒子优胜劣汰等扩大搜索范围.提高获得最优解的概率.实验结果表明服务分配的并行优化及其DPSO求解算法是提高任务工作流执行成功率和时效性的有效方法.Towards how to improve the efficiency and successful probability of task-workflows in service oriented military information system, a parallel optimization method of service resource allocation was proposed. Firstly, a service resource allocation framework is offered. By analyzing the relationship of service executing number, task executing time, task successful probability and service executing cost, a target programming mathematical model for parallel optimization of service resources was established. Then, an ameliorated particle swarm optimization （called DPSO） algorithm was proposed to resolve the matbematical model. By introducing random disturbance, searching boundary optimization and survival of the fittest for the particles, DPSO extended searching scope to obtain the optimal solution with a higher probability. Experimental results show that parallel optimization method of service resource allocation and the DPSO algorithm are effective methods to improve the efficiency and successful probability of task-workflows.

关 键 词：服务并行优化 军事信息服务 粒子群算法 

分 类 号：E89[军事]