Moving Average Single Exponential Smoothing
∑ Ai Ft = Ft−1 + α(At−1 − Ft−1)
Ft = M An =
n
Weighted Moving Average
∑
Ft = M An = wi Ai
Trend-Adjusted Exponential Smoothing
The formula for SEST for the next period, t + 1, can be written as:
SESTt = SFt−1 + Tt−1
where
SFt−1 = pr e v i ou s p er i o d′s f or eca s t + s m o oth ed er r or
Tt = pr e v i ou s p er i o d′s t r e n d + s m o oth ed er r or on t r e n d
SFt−1 = Ft−1 + α(At−1 − Ft−1) = single exponential smoothing
Tt = Tt−1 + β(Ft − Ft−1 − Tt−1) = trend
Forecasting Techniques based on Linear Regression
y = a + bx
n( ∑ x y) − ( ∑ x)( ∑ y) ∑ y − b∑ x
b= a=
n( ∑ x 2) − ( ∑ x)2 n
Linear Regression as a Trend Line
y = a + bt
, Accuracy of Forecasts
Er ror = Act u al − Foreca st
∑ (Act u al − Foreca st)
MAD = = the average absolute error
n
∑ (Act u al − Foreca st)
M A PE = = the absolute error as a percentage of actual value
∑ Act u al
(1e actual waarde neem je niet mee)
Forecast Control
Tracking signal measures whether forecasts keep pace with up-and-down changes in actual
values.
∑ (Act u al − Foreca st)
Track ing sign al =
MAD
Cost-Profit-Volume (CPV) Analysis = Break-even point
Pr of it = Re venu e (R) − Total cost (TC )
Re venu e = Unit Pr ice ( p) × Qu a nt it y (Q)
Total cost = Fi xed cost (FC ) + Var i a ble cost (VC )
Var i a ble cost = var i a ble cost per u nit (v) × Qu a nt it y (Q)
Prof it = ( p − v)Q − FC
Most d esir a ble outcom e
Rel at ive scor e = (wanneer most desirable zo laag mogelijk is)
E valu ated outcom e
E valu ated outcom e
Rel at ive scor e = (wanneer most desirable zo hoog mogelijk is)
Most d esir a ble outcom e
Minimizing distance and costs
∑
Minimize TC = DijWijCij
ij
D = Distance
W = Inderpartmental traffic
C = Cost
ij = between departments i and j
n! = tota al a antal m ogeli jk heden bi j n a antal a fdelingen
∑ Ai Ft = Ft−1 + α(At−1 − Ft−1)
Ft = M An =
n
Weighted Moving Average
∑
Ft = M An = wi Ai
Trend-Adjusted Exponential Smoothing
The formula for SEST for the next period, t + 1, can be written as:
SESTt = SFt−1 + Tt−1
where
SFt−1 = pr e v i ou s p er i o d′s f or eca s t + s m o oth ed er r or
Tt = pr e v i ou s p er i o d′s t r e n d + s m o oth ed er r or on t r e n d
SFt−1 = Ft−1 + α(At−1 − Ft−1) = single exponential smoothing
Tt = Tt−1 + β(Ft − Ft−1 − Tt−1) = trend
Forecasting Techniques based on Linear Regression
y = a + bx
n( ∑ x y) − ( ∑ x)( ∑ y) ∑ y − b∑ x
b= a=
n( ∑ x 2) − ( ∑ x)2 n
Linear Regression as a Trend Line
y = a + bt
, Accuracy of Forecasts
Er ror = Act u al − Foreca st
∑ (Act u al − Foreca st)
MAD = = the average absolute error
n
∑ (Act u al − Foreca st)
M A PE = = the absolute error as a percentage of actual value
∑ Act u al
(1e actual waarde neem je niet mee)
Forecast Control
Tracking signal measures whether forecasts keep pace with up-and-down changes in actual
values.
∑ (Act u al − Foreca st)
Track ing sign al =
MAD
Cost-Profit-Volume (CPV) Analysis = Break-even point
Pr of it = Re venu e (R) − Total cost (TC )
Re venu e = Unit Pr ice ( p) × Qu a nt it y (Q)
Total cost = Fi xed cost (FC ) + Var i a ble cost (VC )
Var i a ble cost = var i a ble cost per u nit (v) × Qu a nt it y (Q)
Prof it = ( p − v)Q − FC
Most d esir a ble outcom e
Rel at ive scor e = (wanneer most desirable zo laag mogelijk is)
E valu ated outcom e
E valu ated outcom e
Rel at ive scor e = (wanneer most desirable zo hoog mogelijk is)
Most d esir a ble outcom e
Minimizing distance and costs
∑
Minimize TC = DijWijCij
ij
D = Distance
W = Inderpartmental traffic
C = Cost
ij = between departments i and j
n! = tota al a antal m ogeli jk heden bi j n a antal a fdelingen
