Sie sind auf Seite 1von 23

Project Scheduling

Probabilistic PERT

PERT Probability Approach to Project Scheduling


Activity completion times are seldom known with cetainty.

PERT is a technique that treats activity completion times as random variables.


Completion time estimates can be estimated using the Three Time Estimate approach. In this approach, three time estimates are required for each activity:
a = an optimistic time to perform the activity P(Finish < a) < .01 m = the most likely time to perform the activity (mode) b = a pessimistic time to perform the activity P(Finish > b) < .01
Results from statistical studies Subjective best estimates

3-Time Estimate Approach Probability Distribution


With three time estimates, the activity completion time can be approximated by a Beta distribution. Beta distributions can come in a variety of shapes:

Mean and Standard Deviation for Activity Completion Times


The best estimate for the mean is a weighted average of the three time estimates with weights 1/6, 4/6, and 1/6 respectively on a, m, and b. Since most of the area is with the range from a to b (b-a), and since most of the area lies 3 standard deviations on either side of the mean (6 standard deviations total), then the standard deviation is approximated by Range/6.
a + 4m + b = the mean completion time = 6 b-a = the standard deviation = 6

PERT Assumptions
Assumption 1
A critical path can be determined by using the mean completion times for the activities. The project mean completion time is determined solely by the completion time of the activities on the critical path.

Assumption 2
There are enough activities on the critical path so that the distribution of the overall project completion time can be approximated by the normal distribution.

Assumption 3
The time to complete one activity is independent of the completion time of any other activity.

The Project Completion Time Distribution


The three assumptions imply that the overall project completion time is normally distributed, with:

= Sum of the s on the critical path


2 = Sum of the 2 s on the critical path

(76 + 4(86) +120)/6

The Probability Approach


(120-76)/6 (7.33)2

Activity Optimistic Most Likely Pessimistic A 76 86 120 B 12 15 18 C 4 5 6 D 15 18 33 E 18 21 24 F 16 26 30 G 10 13 22 H 24 18 32 I 22 27 50 J 38 43 60

90 15 5 20 20 21 21 25 25 14 14 28 28 28 30 30 30 45 45 45

7.33 1.00 0.33 3.00 3.00 1.00 1.00 2.33 2.33 2.00 2.00 1.33 1.33 1.33 4.67 4.67 4.67 3.67 3.67 3.67

2 2 2 53.73 1.00 1.00 0.11 0.11 9.00 9.00 1.00 1.00 1.00 5.43 5.43 5.43 4.00 4.00 4.00 1.77 1.77 1.77 21.81 21.81 21.81 13.47 13.47 13.47

Distribution For Klone Computers


The project has a normal distribution. The critical path is A-F-G-D-J.
A F G D J 90 25 14 20 45 194
2 2 2 2 2 2 A F G D J

53.78 5.44 4 9 13.44

85.66

2 85.66
9.255

Standard Probability Questions


1.

What is the probability the project will be finished within 194 days?
P(X < 194)

2.

Give an interval within which we are 95% sure of completing the project.
X values, xL, the lower confidnce limit, and xU, the upper confidnce limit, such that P(X<xL) = .025 and P(X>xU) = .025

3.

What is the probability the project will be completed within 180 days?
P(X < 180)

4.

What is the probability the project will take longer than 210 days.
P(X > 210)

5.

By what time are we 99% sure of completing the project?


X value such that P(X < x) = .99

Excel Solutions
NORMDIST(194, 194, 9.255, TRUE) NORMINV(.025, 194, 9.255) NORMINV(.975, 194, 9.255) NORMDIST(180, 194, 9.255, TRUE)

1 - NORMDIST(210, 194, 9.255, TRUE)

NORMINV(.99, 194, 9.255)

Using the PERT-CPM Template for Probabilistic Models


Instead of calculating and by hand, the Excel template may be used. Instead of entering data in the and columns, input the estimates for a, m , and b into columns C, D, and E.
The template does all the required calculations After the problem has been solved, probability analyses may be performed.

Enter a, m, b instead of

Call Solver Click Solve Go to PERT OUTPUT worksheet

Call Solver Click Solve

To get a cumulative probability, enter a number here

P(Project is completed in less than 180 days)

Cost Analysis Using the Expected Value Approach


Spending extra money, in general should decrease project duration. But is this operation cost effective? The expected value criterion can be used as a guide for answering this question.

Cost Analyses Using Probabilities


Suppose an analysis of the competition indicated:
If the project is completed within 180 days, this would yields an additional profit of $1 million. If the project is completed in 180 days to 200 days, this would yield an additional profit of $400,000.

KLONE COMPUTERS Cost analysis using probabilities


Completion time reduction can be achieved by additional training. Two possible activities are being considered.
Sales personnel training: (Activity H)
Cost $200,000; New time estimates are a = 19, m= 21, and b = 23 days.

Technical staff training: (Activity F)


Cost $250,000; New time estimates are a = 12, m = 14, and b = 16.

Which, if either option, should be pursued?

Analysis of Additional Sales Personnel Training


Sales personnel training (Activity H) is not a critical activity.
Thus any reduction in Activity H will not affect the critical path and hence the distribution of the project completion time.

This option should not be pursued at any cost.

Analysis of Additional Technical Staff Training


Technical Staff Training (Activity F) is on the critical path so this option should be analyzed. One of three things will happen:
The project will finish within 180 days:
Klonepalm will net an additional $1 million

The project will finish in the period from 180 to 200 days
Klonepalm will net an additional $400,000

The project will take longer than 200 days


Klonepalm will not make any additional profit.

The Expected Value Approach


Find the P(X < 180), P(180 < X < 200), and P(X > 200) under the scenarios that
No additional staff training is done Additional staff is done

For each scenario find the expected profit: Expected Additional Profit
1000000(P(X<180)) + 400000(P(180<X<200)) + 0(P(X>200))

Subtract the two expected values. If the difference is less than the cost of the training, do not perform the additional training.
Caution: These are expected values (long run average values). But this approach serves as a good indicator for the decision maker to consider.

The Calculations
The PERT-CPM template can be used to calculate the probabilities.
No Additional Training = 194 = 9.255 .065192 $ 65,192 .676398 $270,559 .258410 .111287 .729561 $291,824 Additional Training = 189 = 9.0185 .159152 $159,152

P(X < 180)


X $1000000

P(180 <X < 200)


X $400000

P(X > 200)


X $0

Total = $335,751

Total = $450,976 Do not perform the additional training!

Net increase = $450,976-$335,751 = $115,225 This is less than the $250,000 required for training.

Review
3-Time Estimate Approach for PERT
Each activity has a Beta distribution Calculation of Mean of each activity Calculation Variance and Standard Deviation for each activity

Assumptions for using PERT approach Distribution of Project CompletionTime


Normal Mean = Sum of means on critical path Variance = Sum of variances on critical path

Using the PERT-CPM template Using PERT in cost analyses

Das könnte Ihnen auch gefallen