用PERT法求解不确定网络中活动时间参数的改进研究

基于计划评审技术(PERT)中β分布的假设,阐述了常规三时估计方法的不足,提出了相应解决对策,即采用限定概率三时估计法分别建立了β分布和Gamma分布拟合方差最小模型,用微粒群算法求解模型,并且将求解结果与常规三时估计方法进行了对比研究。实例证实,对于不同类型的活动,可以选择合理的分布类型并利用限定概率三时估计法,对各个活动的持续时间服从的分布密度函数进行拟合,从而提高PERT的精度和效率。

To Study on the Improving with PERT Finding Solution to Time Parameters of Activities in the Uncertain Network

Although Program Evaluation and Review Technique(PERT) and Monte Carlo Simulation(MCS) used in computing PERT network has been widely applied in project,some deficiencies still exist.This paper builds upon the beta distribution hypothesis,points out the deficiencies of traditional three-time estimation method,provides solution that is building the minimum fitting variance models of both beta distribution model and gamma distribution by using the limited probability three-time estimation method,using particle swarm optimization(PSO) to solve the models,and compare the results from fitting method and traditional method.A case shows that for different types of activities,to choose the reasonable distribution style and to use limited probability three-time estimation method to fitting the distribution of each activity can improve the accuracy and efficiency of PERT.