基于粒子群算法的多项目资源均衡方法研究  

The research for the multi-project resource equalized methods based on particle swarm optimization

作　　者：王秋全 李向[2] 王岭玲 

机构地区：[1]中国交通建设股份有限公司西北区域总部,陕西西安710065 [2]中国地质大学计算机学院,湖北武汉430074 [3]武汉工程科技学院机械与电子信息学部,湖北武汉430200

出　　处：《应用科技》2014年第3期55-59,共5页Applied Science and Technology

基　　金：中国地质大学(武汉)横向科研基金资助项目(2012196539);湖北省自然科学基金资助项目(2012195075)

摘　　要：针对传统粒子群算法容易陷入局部最优的缺点,提出利用动态惯性权重参数和模拟退火算法修改突变概率,进而改进传统粒子群算法,探讨各项目工期最短情况下的多项目资源均衡分配问题。通过对比试验表明,改进的粒子群优化(particle swarm optimization,PSO)算法很好地实现了多项目的资源均衡优化,通过同比试验验证了改进PSO算法在解决不同规模多项目的资源均衡问题时的算法时间复杂度的线性增长性,很好地表达了人们的调度意图。For the traditional particle swarm optimization easily falling into local optimum , this paper proposes a method that dynamic inertia weight parameter and simulated annealing algorithm are applied to modify the mutation probability to improve the traditional particle swarm optimization , and investigates the balanced allocation problem of multi-project resource under the circumstances that each project has the shortest duration .The comparative exper-iments indicate that the improved particle swarm optimization can nicely achieve the resource balanced optimization of multi-project, and year-on-year experiments verify the linear growth of time complexity of the improved PSO when solving resource balanced optimization of multi-project in different sizes , which can better express the inten-tion of scheduling .

关 键 词：粒子群优化 调度网络 均衡优化 多项目 

分 类 号：TP35[自动化与计算机技术—计算机系统结构]