运筹学与供应链管理-第2讲

3-1McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting第２讲需求预测3-2McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting预测:•是关于未来的描述•预测用于帮助管理人员–系统计划：产品服务的类型、设施配置、厂址选择–实施计划：库存、劳动力、采购、生产进度3-3McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting会计成本/利润估计财务现金流量人力资源招工/招聘/培训营销价格,促销,战略管理信息系统服务运作调度,MRP,工作负荷产品/服务设计新产品和服务预测的用途3-4McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting•假定系统中存在因果关系过去==未来•由于不确定性因素的存在使得预测很少准确无误•对一组事物的预测比对单个事物预测准确•当预测时间跨度增加时，预测精度将下降我看这学期你能得优秀成绩。3-5McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting好的预测方法的基本要素适时精度可靠书面3-6McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting预测进行的步骤步骤1确定预测目的步骤2确定预测时间跨度步骤3选择预测技术步骤4收集和分析数据步骤5准备预测步骤6预测监控“预测”3-7McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting预测的类型•主观判断–主观意见•时间序列–使用历史数据，认为将来和过去相似•联合模型–基于自变量预测未来3-8McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting主观预测•经理人员的意见•与顾客直接接触人员的意见•消费者调查•其它预测方法–德尔非法Delphimethod3-9McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting时间序列预测•长期趋势–数据的长期移动•季节性变动–数据短期规则变化•不规则变动–异常情况引起的变动•随机变动–各种可能性引起的变动3-10McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.ForecastingCh10-7©2000byPrentice-HallIncRussell/TaylorOperMgt3/e需求变动的形式需求时间趋势变动随机变动需求时间季节变动需求时间需求时间周期变动带季节性的趋势变动3-11McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting•简单易用•基本上不需要成本•不需要数据分析•容易理解•不能提供较高的精度•能够作为衡量精度的标准简单预测法3-12McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting•平稳时间序列数据–F(t)=A(t-1)•季节变动–F(t)=A(t-n)•长期趋势数据–F(t)=A(t-1)+(A(t-1)–A(t-2))简单预测法应用3-13McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting简单预测唉,给我点时间....上周我们卖了250轮胎....那么,下周我们应该卖....3-14McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting平均方法•移动平均•加权移动平均•指数平滑法3-15McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting简单移动平均图3-4MAn=nAii=1n35373941434547123456789101112实际数据MA3MA53-16McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.ForecastingCh10-14©2000byPrentice-HallIncRussell/TaylorOperMgt3/e加权移动平均•可调整和反映简单移动平均法中不同时期数据的影响WMAn=i=1WiDi此处,Wi=第i期的权重，其百分值在0~100之间Wi=1.003-17McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.ForecastingCh10-15©2000byPrentice-HallIncRussell/TaylorOperMgt3/e加权移动平均之例月份权重数据八月17%130九月33%110十月50%90十一月的预测313050900331100171301034WMAWDiii(.)()(.)()(.)().3-18McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting指数平滑法假设—最近的观测具有最高的预测价值.–因此在预测时，应该给更近的数据赋予更高的权重.Ft=Ft-1+(At-1-Ft-1)3-19McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting时期实际数据Alpha=0.1误差Alpha=0.4误差14224042-2.0042-234341.81.2041.21.844041.92-1.9241.92-1.9254141.73-0.7341.15-0.1563941.66-2.6641.09-2.0974641.394.6140.255.7584441.852.1542.551.4594542.072.9343.131.87103842.36-4.3643.88-5.88114041.92-1.9241.53-1.531241.7340.92指数平滑法举例3-20McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting选择平滑常数35373941434547123456789101112PeriodDemand.1.4实际数据3-21McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting线性趋势方程•b＝直线斜率Yt=a+bt012345tYb=n(ty)-tynt2-(t)2a=y-btn3-22McGraw-Hill/IrwinOperationsManagement,SeventhEdition,byWilliamJ.StevensonCopyright©2002byTheMcGraw-HillCompanies,Inc.Allrightsreserved.Forecasting线性趋势方程举例ty周t2销售ty111501502415731439162486416166664525177885t=15t2=55y=812ty=2499(t)2=2253-23McGraw-Hill/IrwinOpera