基于蒙特卡洛方法的网络计划资源优化  

Resource Optimization for Network Plan Based on Monte Carlo Method

作　　者：张厚先[1] 施柏楠 

机构地区：[1]南京工程学院建筑工程学院 [2]四川大学建筑与环境学院

出　　处：《数学的实践与认识》2015年第6期120-127,共8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：蒙特卡洛方法由计算机产生伪随机数而生成试验点,根据约束条件找到优化解,据此编制C语言程序,可以成功求解网络计划的工期固定-资源均衡优化、资源有限-工期最短优化和资源有限-工期最短-资源均衡优化,并给出蒙特卡洛方法得到最优解的概率.算例结果表明:相同条件下,基于蒙特卡洛方法的工期固定-资源均衡优化方案资源方差较粒子群算法小,基于蒙特卡洛方法的资源有限-工期最短优化方案工期较遗传算法短,基于蒙特卡洛方法的资源有限-工期最短-资源均衡优化方案工期较遗传算法短.Monte Carlo method can generate test points by a computer random number generator to find the optimized solution based on the constraints.Duration fixed-resource leveling optimization,limited resources-the shortest period optimization and limited resourcesthe shortest period-resource leveling optimization were successfully solved based on Monte Carlo method through the C language program.Probability to obtain optimized solution was given.Numerical results showed that：：under the same conditions,the variance of the duration fixed-resource leveling optimization program based on the Monte Carlo method was smaller than the particle swarm,the duration of limited resources-the shortest period optimization program based on the Monte Carlo method was shorter than the genetic algorithm,the duration of limited resources-the shortest period-resource leveling optimization program based on the Monte Carlo method was shorter than the genetic algorithm.

关 键 词：网络计划 资源优化 蒙特卡洛方法 

分 类 号：O242.2[理学—计算数学]