
214 RFID and Auto-ID in Planning and Logistics
17.7 METHODS FOR FORECASTING STATIONARY SERIES
Two popular techniques for forecasting stationary time series are moving averages and exponential
smoothing. A stationary time series is a time series where each observation can be represented by a
constant and a random uctuation. In symbols,
D
t t
= +μ ε
where
μ is an unknown constant corresponding to the mean of the series
ε
t
is a random error with mean zero and variance σ
2
The exponential and moving average methods are more precisely known as single or simple
exponential smoothing and single or simple moving averages. Since the underlying model assumed ...