Scribd
Upload
11 views

Characteristics of A Project

Uploaded by

panyexing
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
11 views

Characteristics of A Project

Uploaded by

panyexing
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 8

Characteristics of a Project

1)Low volume/Unique – not repetitive

2)Non-standardized inter-dependent activities

3)Timebound, temporary – Clear start and end

4)Specific, clear goals – time, cost, specs (may change


though)

5)But outcomes are uncertain

6)Teams – banding and disbanding

7)Conflict

As contrasted to a Process: High volume, standardized


activities and outcomes, routine part of business (not
temporary)
WBS – work break down structure
GANTT CHART covers sequence, activity time, activity
dependency project completion time, cost.(risk not
covered)
CPM- Critical path method
Slack – Latest start time – Earliest start time
Slack = 0 are critical activity.
Latest start time + Activity time = Latest completion time

Forecast
Naïve forecast : F = A t t−1

Cannot provide high accuracy


But, Can be a standard for accuracy
前一个 period 的 demand 作为 forecast
F t= A t−1 +T T is trend value
Cumulative method: Assumes all past data is equally
important in predicting the future. demand level (L)
• At = L+et
F t=
∑ of period of time∈ previous
time frame

The moving average: Let the forecast equal the average


of the last N observations
Forecast demand =total demand in Period of number /
Period
• If N is small, the forecast will quickly respond to
“real” changes but might respond to noise.
• If N is large, the forecast is unlikely to respond to
noise, but will be slow to respond to “real” changes
in the system.
Weights: sum of demand * weight in period of
time(percentage)
Or sum of demand * weights in period of time / sum of
weights
Single exponential smoothing:
Ft+1 = α Dt + (1 – α) Ft, where:
Ft+1 = Forecast for tomorrow
Dt = Actual demand recorded today
Ft = Demand that was forecasted for today
α = Alpha, a weight that ranges between 0 and 1.00,
also called a ‘smoothing constant.’

 Forecast error = Actual minus forecast


 Forecast accuracy % = (100 - % forecast error) = [100
– (Forecast error/Actual)*100]
 Cumulative forecast error (CFE) = Sum of all errors.
 Average (mean) error = CFE/Number of periods
Mean Error(ME) and (MP, percentage E): Negative errors
and positive errors can cancel each other out, therefore,
this is a good measure of bias, but not of accuracy.
ME=sum of (A-F)/n
Mean absolute error (MAE):If we take the absolute value
of each error term,we will have a measure that better
shows the accuracy of the forecast.
MAE = =sum of |A-F|/n: Less intuitive than MAE
Large deviations have a larger impact than small
deviations , Easier computationally
** APE = Absolute % error = (Absolute Error/Actual) *100 = e.g. Fall 1 at α:0.2
= (3700/5700) * 100 ≈ 65%
*** AD = Absolute deviation = |Error|

Root mean square error(RMSE): Standard deviation of


the error
MAPE (confident interval): As a very general rule, a MAPE
of 10%-20% is considered good, but for some
applications, 40% might be good; for others, it might be
possible to obtain a MAPE of 3%. (Can’t calculate if actual
demand is ever zero. A huge error in a low volume can
skew the results.)
CFE = sum of errors
 ( At  Ft ) 2
U
 ( At  At  1 ) 2
Theil’s U-statistic:
• If U=1: the naïve method is as good as the
forecasting technique being used.
• If U<1: the forecasting technique being used is better
than the naïve method. The smaller U, the better
the technique.
• If U>1: the naïve method produces better results
than the technique being used.

• CVD = standard deviation / mean. If CVD is high


(>0.8 is one possible benchmark), we can’t expect
very accurate forecasts.
R SQUARE number is percentage around mean.
Higher the better.
Significance F: confidence level of (1-F).

P-value: The P-value for advertising reads as


0.003747, well under the stipulated 0.05 standard,
indicating a confidence level of 99.6253% (1 minus P-
value). However, the P-value for Price is 0.230758, is
much above the required 0.05 mark, indicating a
confidence level of only 76.69242% (1 minus P-
value). Compare with 95% minimum confidence
level is generally prescribed for business models

EOQ = sqrt( 2*D*S /H)


D = Demand per year
S = Setup (order) cost per order
H = Holding (carrying) cost
d = Demand per day
L = Lead time in days

Expected number of order = D/Q(eoq)


Expected time between order = working day / yr / N

Inventory turnover = 1/ flow time


Also = COGS / Average aggregate inventory
Internal : average aggregate inventory
Average days of inventory on hand
Inventory turnover
Customer facing:
Average CSL
Either 100% or 0%
Average fill rate
Fill rate = inventory / demand

Responsiveness: inventory elasticity

You might also like