基于粒子群算法的大型网络计划费用优化  

Cost Optimization for a Large Network Plan Based on Particle Swarm Optimization

作　　者：张厚先[1] 杨昭兰[2] 

机构地区：[1]南京工程学院建筑工程学院,江苏南京211167 [2]南京工程学院工业中心,江苏南京211167

出　　处：《数学的实践与认识》2015年第11期142-148,共7页Mathematics in Practice and Theory

基　　金：住房和城乡建设部科学技术计划项目2014-K3-039

摘　　要：大型网络计划费用优化对科学有效地进行工程项目进度管理具有重要意义,但大型网络计划费用优化随工作个数增加而约束方程和计算量骤增,成为数学和计算机科学领域至今未解决的难题.借助建立评价函数、设计进化方程、设计网络计划时间参数的计算机算法等基础工作,选择工作持续时间作为粒子空间坐标并设置可行解范围,用蒙特卡洛方法和限制条件优化初始粒子群,用二维动态数组解决大型网络计划粒子群算法优化运行image超限问题,成功求解有61个工作的大型网络计划费用优化算例.因此,经过特定设计的粒子群算法是微机和有限的计算时间条件下求解大型网络计划费用优化问题的一个有效方法.Cost optimization for a large-scale network plan is important to manage the progress of the project effectively and scientifically. But cost optimization for a large-scale network plan become problems unresolved so far in mathematics and computer science be- cause constraint equations and the number of calculations surge while the number of job increasing. Based on establishing the evaluation function, designing evolution equations, and designing computer algorithm for parameters of the network plan, etc., select durations as the spatial position of the particle, set feasible solution range, optimize the initial particle swarm optimization using Monte Carlo method and restrictions, solve the image overrun problem for large-scale network plan optimization using two-dimensional dynamic array, cost optimization for a large-scale network plan with 61 works has been solved successfully. Therefore, the specially designed particle swarm optimization algorithm is an effective way to solve the cost optimization problems for a large-scale network plan under conditions of limited computing time by a microcomputer.

关 键 词：大型网络计划 费用优化 粒子群算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]