SMA:
Single Moving Average
Simple moving average is a method simply by adding the past demands in a set of time periods ( e.g. three days ) and then dividing this total by time periods of this set ( e.g. three days ).
The advantages of this method are quite general, one is that SMA is a mathematical model, therefore it is objective and logical, and as the formula is simple, it is very easy for using.
There are many cases that the managers need to forecasting the short-term demand for a large number of different items, the cost of developing specific forecast for each individual products would be considerably high in terms of time and money, therefore a method which can be applied easily on several products and come out with a reasonably precise prediction would be essential. The SMA is a very good technique for short-term forecasting, since the short-term averages respond quickly to changes in the demand of the underlying. (Lancaster G and Lomas R, 1985)
However, there are few limitation by using this method, one is that the data collected have to be stationary, otherwise SMA would attempt to smoothing the seasonal peak, and tend to horizontal the data (Lancaster G and Lomas R, 1985). In addition, SMA is not considering any causal relationship within the data.
In general , SMA is used when the demand is not in a rapid increase or decrease trend, and without any seasonal factors that possible to affect the forecast.
Example:
3-Weeks SMA:
3-Weeks SMA:
F4 = ( 35 + 32 + 26 ) / 3 = 31
F5 = ( 32 + 26 + 28 ) / 3 = 28.6 ≈ 29
F6 = ( 26 + 28 + 34 ) / 3 = 29.3 ≈ 30
. . .
The word ‘Moving’ means that as new values become available, the new data must replace the oldest data. Thus, the data set is constantly "Moving" to account for new data as it becomes available.
In addition, there is a method call ‘ Second Order Simple Moving Averages’, which can not use in forecast directly, but it can provide a smooth trend of forecasted demand, it is using the same method as SMA, but instead of using the actual demand data, ‘second order SMA’ uses the forecasted data from SMA:
Copyright © 2015 | University of Warwick, Coventry | [email protected]
POWERED BY KASPAR CHEN