The Forecast in SOP 在SOP中的预测

来源：http://help.sap.com/printdocu/core/Print46c/en/data/pdf/LOPR/LOPR.pdf


The Forecast in SOP在SOP中的预测

With the forecast, you can estimate the future progression of values in a time series on the basis of past history. You do this either online or in the background.In standard SOP, you can forecast the sales quantities of a product group or material. The system bases the forecast on the historical consumption of materials. It then aggregates theseresults to the product group level. Consumption data includes every kind of goods issue, even goods that have been written off as scrap.
通过SAP的预测，你可以依据过去的历史记录来估算未来值。这些工作可以在线或是后台运行。在标准的SOP中，你可以预测物料或产品组的销售数量。系统计算物料的消耗历史，然后汇总这些结果至产品组的层次。消耗数据包括多种发料，甚至包括了报废。

In flexible planning, you can forecast any key figure that you want, provided that it has been defined for forecasting in Customizing (in Set parameters for info structures and key figures). The system bases the forecast on the actual values of this key figure. In level-by-level planning, you can also base the forecast on historical consumption data. In this case, the information structure must contain the characteristics “material” and “plant,” and you must set the Consumpn. indicator in the forecast profile.


在柔性计划中，你可以预测任何已定义在预测中的关键指标（对于信息结构和关键指标设定的参数）。系统基于这些关键指标的实际值来做预测。在逐层计划中，你可以基于历史消耗数据进行预测。在此情形下，信息结构中必须包括特征“物料”和“工厂”，并且你必须在预测参数文件中设置消耗指示符。

In level-by-level planning, you can base the forecast on the consumption quantities of a
“reference material.” This is useful if you do not have any historical values for a material; for example, because it is new. You define the reference material in the material master record.
在逐层计划中，你可以使用“参考物料”的历史消耗数量来做预测。这在你创建一个新物料，没有任何历史消耗数据时非常有用。其定义在物料主数据中。（MM02)


The historical data is automatically aggregated to the current planning level before the forecast is carried out. If you run the forecast online, this is the planning level on which you are working in the planning table. If you run the forecast in the background, this is the planning level you define in the job variant.
 在预测完成前，历史数据会自动汇总到当前的计划层次。如果在线进行预测，计划的层次是你现在操作的计划表。如果是后台运行，则取决于你定义的作业变式。

  If you run the forecast at a high level and the results are disaggregated to the
detailed level, the results will be different than if you carry out the forecast at the
detailed level.
如果你在高层次运行预测，并且结果分散到细节层，则这个结果可能与你在细节层面做的预测存在差异。



Forecast Models 预测模型   
     When a series of consumption values is analyzed, it normally reveals a pattern or patterns.  These patterns can then be matched up with one of the forecast models listed below:

1. Constant ：consumption values vary very little from a stable mean value
2. Trend ：consumption values fall or rise constantly over a long period of time with only occasional deviations 
3. Seasonal： periodically recurring peak or low values differ significantly from a stable mean value 
4. Seasonal trend ： continual increase or decrease in the mean value
5. Copy of actual data ：(no forecast is executed)copies the historical data updated from the operative application, which you can then edit
6. Irregular： no pattern can be detected in a series of historical consumption values 

 当对一系列的消耗数据进行分析时，通常是使用一些规则或模式。这些模式如下：

1. 常量： 消耗值在一个稳定值左右，只有较小的偏离。
2. 趋势 ：在一个较长的时期内，消耗值规律性的上升下降，只会偶尔有大的偏差。
3. 季节： 周期性的循环呈现高峰与低谷，区别于稳定在一个平均值。
4. 季节趋势：持续性的呈现上升或下降趋势
5. 复制实际数据：从实际业务复制消耗数据，然后编辑这些数据（不执行预测）
6. 不规则：对于历史数据没有任何规则可循。

Model Selection
Before you run a forecast, you must specify which model the system should use to calculate the forecast values.
There are three possibilities:
  Manual model selection
  Automatic model selection
  Manual model selection with the system also testing for a pattern

模型选择：
在运行预测前，你必须指定一个模型来让系统计划预测值。有下列三种选择：

1. 手动模型选择
2. 自动模型选择
3. 带系统测试规则的手动模型选择

Manual Model Selection
If you want to select a model manually, you must first analyze the historical data to determine whether a distinct pattern or trend exists. You then define your forecast model accordingly.

手工模型选择
如果你想手工选择预测模型，你需要先对历史数据进行分析，再来确定是否有清晰的模式或趋势存在，然后依此确定预测模型。

Constant pattern
常量规则
If your historical data represents a constant consumption flow, you select either the constant model or the constant model with adaptation of the smoothing factors. In both cases, the forecast is carried out using first-order exponential smoothing. When adapting the smoothing parameters,the system calculates different parameter combinations and then selects the optimum parameter combination. The optimum parameter combination is the one which results in the lowest mean absolute deviation.
如果历史数据呈现出常量的消耗形势，你可以选常量模型或是使用带平滑因子的常量模型。在上述两种情况下，预测都是使用一阶指数平滑来完成的。当获得平滑因子后，系统计算不同参数的组合，然后选择参数的最佳组合。最佳参数组合就是与实际偏差最小的那个结果。

You have another two possibilities if the historical pattern is constant; either the moving average model or the weighted moving average model. In the weighted moving average model, you weight individual historical values with the result that the system does not give equal value to historical data when calculating the forecast values. By so doing, you can influence the calculation so that more recent historical values play a more important role in the forecast than less recent ones as is also the case with exponential smoothing.
如果历史(消耗数据的）规律是常量型，则还可又使用移动平均模型和权重移动平均模型。在权重移动平均模型中，可对单个历史值指定权重。对于越是靠近现在的数据给它更高的权重，使它在计算中的地位更重要。

Trend pattern
If your historical data represents a trend, you should select either the trend model or a second-order exponential smoothing model. In the trend model, the system calculates the forecast values using first-order exponential smoothing. In the second-order exponential smoothing models, you can choose a model with or without a model  parameter optimization.

趋势规则
    如果历史数据呈现趋势，你可以选择趋势模型或二阶指数据平滑模型。 在趋势模型中，系统使用一阶指数据平滑来计算预测值。 在二阶指数据平滑模型中，你可选择使用或不使用模型参数优化。

Seasonal pattern
If your historical data represents a seasonal pattern, you specify the seasonal model. The system calculates the forecast values for the seasonal model using first-order exponential smoothing.
季节规则
如果历史数据呈现季节性规律，你使用季节模型。系统计算其预测值使用一阶指数据平滑。

Seasonal Trend pattern
If your historical data represents a seasonal trend pattern, you select a seasonal trend model.The system calculates the forecast values using first-order exponential smoothing.
季节趋势规则
如果历史数据呈现季节趋势规律，你可以选择季节趋势模型。系统计算其预测值使用一阶指数据平滑。

Irregular pattern
If you cannot detect any of the above trends or patterns, and you still want the system to carry out a forecast, it is usually advisable to select either the moving average model or the weighted moving average model.

不规则的规律
如果不适用以上任何规则，但你仍希望系统完成预测时，则建议使用移动平均和权重移动平均模型。