l9 Dynamics 4

ESD.
36J System & Project Management

+
Dynamics of Project Performance
System Dynamics and Project Management Class Four (10/2/03)
Copyright 2003 James M. Lyneis
Topics
Review Practice for 10/3 Staff and schedule dynamics Managing the Dynamics & Summary Homework 4
10/02/03
- ESD.36J SPM
Productivity Minimum Time to Perform a Task
Normal Quality Quality
Potential Work Rate Maximum Work Rate Feasible Work Rate
Quality on Quality Dynamics

Staff Level
Work Accomplishment Table for Effect of Prior Work Quality on Quality Effect of Prior Work Quality on Quality Work to Do Rework Generation <Initial Work to Do> Undiscovered Rework Work Done Project Finished Switch
Average Work Quality Maximum Time to Discover Rework
<Rework Generation> Rework Discovery Rate of Doing Work <Work Accomplishment> Cumulative Work Done
<Work Done>
Time to Discover Rework Minimum Time to Discover Rework
Work Believed to Be Done Fraction Really Complete Effort Expended Fraction Perceived to be Complete Effect of Work Progress <Staff Level> <Initial Work to Do> Table for Effect of Work Progress
Cumulative Effort Expended
<Project Finished Switch>
10/02/03
- ESD.36J SPM
+
1
Graph for Quality
0.85
0.7
0.55
0.4
12
15
18
21
24
27 30 33 Time (Month)
36
39
42
45
48
51
54
57
60
Quality : Class3 Step1 Quality : Class2 Step2
Fraction Fraction
10/02/03
- ESD.36J SPM
+
1
Graph for Effect of Prior Work Quality on Quality
0.85
0.7
0.55
0.4
12
15
18
21
24
27 30 33 Time (Month)
36
39
42
45
48
51
54
57
60
Effect of Prior Work Quality on Quality : Class3 Step1 Effect of Prior Work Quality on Quality : Class2 Step2
Dimensionless Dimensionless
10/02/03
- ESD.36J SPM
+
1 Fraction 100 Task
Quality
0.75 Fraction 75 Task
0 Fraction 0 Task 0
12
15
18
21
24
27 30 33 36 Time (Month)
39
42
45
48
51
54
57
60
Average Work Quality : Class3 Step1 Quality : Class3 Step1 Undiscovered Rework : Class3 Step1 Work Done : Class3 Step1
Fraction Fraction Task Task
10/02/03
- ESD.36J SPM
+
100
Graph for Work Done
75
50
25
0
0 3 6 9 12 15 18 21 24 27 30 33 Time (Month) 36 39 42 45 48 51 54 57 60 Task Task
Work Done : Class3 Step1 Work Done : Class2 Step2
10/02/03
- ESD.36J SPM
With short time to discover rework, quality is better ...

Graph for Quality
0.85
0.7
0.55
0.4 0 3 6 9 12 15 18 21 24 27 30 33 Time (Month) 36 39 42 45 48 51 54 57 60
Quality : Class3 Step1 Quality : Class3 Step1 Short TDRW
Fraction Fraction
10/02/03
- ESD.36J SPM
20
because undiscovered rework is lower

Graph for Undiscovered Rework
15
10
0 0 3 6 9 12 15 18 21 24 27 30 33 Time (Month) 36 39 42 45 48 51 54 57 60 Task Task
Undiscovered Rework : Class3 Step1 Undiscovered Rework : Class3 Step1 Short TDRW
10/02/03
- ESD.36J SPM
200
Less work needs to be done and the project finishes sooner

Graph for Cumulative Work Done
150
100
50
0 0 3 6 9 12 15 18 21 24 27 30 33 Time (Month) 36 39 42 45 48 51 54 57 60 Tasks Tasks
Cumulative Work Done : Class3 Step1 Cumulative Work Done : Class3 Step1 Short TDRW
10/02/03
- ESD.36J SPM
10
Lesson: Its the undiscovered rework!
10/02/03
- ESD.36J SPM
11
Topics
10/02/03
- ESD.36J SPM
12
Staff and Schedule Dynamics

Schedule pressure Adding staff Slipping schedule
10/02/03
- ESD.36J SPM
13
Schedule Pressure
10/02/03
- ESD.36J SPM
14
Schedule Pressure
How might management respond when a project falls behind schedule?
10/02/03
- ESD.36J SPM
15
Schedule Pressure

Pressure team to work faster Work longer hours/overtime
10/02/03
- ESD.36J SPM
16
Schedule Pressure

Pressure team to work faster Work longer hours/overtime Haste makes waste Fatigue adds to mistakes (and may reduce productivity)
Downsides -
10/02/03
- ESD.36J SPM
17
Table for Effect of Schedule Pressure on PDY Effect of Schedule Pressure on Productivity
Anticipated Schedule Overrun
<Project Finished>
Scheduled Completion Date Normal Productivity Initial Scheduled Completion
Perceived Real Completion Date
Productivity
<Time> Indicated Completion Date Based on Progress
<Project Finished>
Staff Level
Estimated Cost to Complete <Work to Do>
10/02/03
- ESD.36J SPM
Average Productivity Copyright 2003 James M. Lyneis
19
Expected completion date:

When will the project finish, given current time, estimated person-years remaining, and staff? Estimated person-years remaining = work to do divided by productivity (person-months) Average Productivity = work believed to be done divided by cumulative effort expended (tasks/person-month)
10/02/03
- ESD.36J SPM
20
Effect of Schedule Pressure ...

On Productivity On Quality
okup - Table for Effect of Schedule Press 1
up - Table for Effect of Schedule Pressur 2
0 -0.2
0 0 1
Anticipated Schedule Overrun (fraction) The tables are the inverse of each other
10/02/03 - ESD.36J SPM Copyright 2003 James M. Lyneis 21
Example
Effect of Schedule Pressure on Productivity = Sensitivity for Effect of Schedule Pressure on Productivity * Table for Effect of Schedule Pressure on Productivity + (1 - Sensitivity for Effect of Schedule Pressure on Productivity)
10/02/03
- ESD.36J SPM
22
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
Output = Input Table * Sensitivity + (1- Sensitivity)

Output of Sensitivity Function
5 Input 0
6 0.5
8 1.5
10
Input Table
10/02/03 - ESD.36J SPM
23
Schedule Pressure Simulations

If the effect of schedule pressure on quality is the exact opposite of that on productivity, schedule pressure in this simple model causes the project to finish somewhat later and cost more (Class3 Step2 vs Step1). However, if the quality effect is stronger than the productivity effect (Class3 Step2 PDY), then schedule pressure causes a significant delay in finishing.
10/02/03
- ESD.36J SPM
24
+
400
Graph for Cumulative Work Done
300
200
100
0
0 3 6 9 12 15 18 21 24 27 30 33 Time (Month) 36 39 42 45 48 51 54 57 60 Tasks Tasks Tasks
Cumulative Work Done : Class3 Step1 Cumulative Work Done : Class3 Step2 Cumulative Work Done : Class3 Step2 PDY
10/02/03
- ESD.36J SPM
25
Lesson: Monitoring/creating the wrong pressures can be counterproductive.
10/02/03
- ESD.36J SPM
26
Staffing
10/02/03
- ESD.36J SPM
27
Adding Staff
Assumptions
hiring/transfer in delays equal 4 months, transfer out 1 month 24 months are required to gain full experience inexperienced staff work at 50% productivity and quality of experienced staff there is no limit on staff available
10/02/03
- ESD.36J SPM
29
Changing Schedule (not included in Class4.mdl)
10/02/03
- ESD.36J SPM
30
Summary of Model to Date

Model has 3 effects on P & Q :

Quality on quality Schedule pressure Experience
Decisions to increase or reduce staffing Decisions to change scheduled completion date (to be added later)
10/02/03
- ESD.36J SPM
31
Topics
10/02/03
- ESD.36J SPM
32
Project Staffing
Now that we understand what causes this behavior ...

... what can we do about it?
Typical Plan
Time
How does it get started?

Time Remaining Skill & Experience Staff Morale Organisational Size Changes Turnover Overtime Hiring
Scheduled Completion Time
Schedule Pressure Productivity Out-of-Sequence Work

Progress
Work To Be Done
Quality
Equivalent Staff on Project Hours Expended to Date
Staffing Requested
Expected Completion Availability of Prerequisites Time
Added Work
Work Really Done
Expected Hours at Completion
Work Quality to Date
Known Rework
Undiscovered Rework
Obsoleted Work
Rework Discovery
Perceived Progress
10/02/03
- ESD.36J SPM
34
How does it get started?
Dynamics are initiated by (1) an infeasible initial plan, or (2) changes adding scope and obsoleting completed work ...
People Productivity Quality Progress
Work To Be Done
Work Really Done
Scope Growth
Known Rework
Rework Discovery
Customer Changes
Undiscovered Rework
10/02/03
- ESD.36J SPM
35
What makes it worse?


Progress
Work To Be Done
Quality
Staffing Requested
Added Work
Work Really Done
Known Rework
Undiscovered Rework
Obsoleted Work
Rework Discovery
Perceived Progress
10/02/03
- ESD.36J SPM
36
Corrective actions to get the project back on track...
People
Productivity Quality
Progress
Overtime Scope Growth
Work To Be Done
Work Really Done
Customer Changes
Known Rework Undiscovered Rework
Rework Discovery
Schedule Acceleration
Hiring
Apparent Progress
10/02/03
- ESD.36J SPM
37

Out-of-Sequence Work, Worksite Congestion, Coordination Problems, Morale Problems
create vicious circles that undercut intended effects

Fatigue, Burnout
Average Employee Skill and Quality
People Productivity Quality
Progress
Overtime
Scope Growth
Work To Be Done
Work Really Done
Customer Changes
Known Rework Undiscovered Rework
Rework Discovery
Schedule Acceleration
Hiring
Apparent Progress
Remember -PRODUCTIVITY = NORMAL PRODUCTIVITY * EFFECT OF STAFF EXPERIENCE * EFFECT OF QUALITY OF PRIOR WORK* ... Dimensions: Productivity -- Tasks/Month/Person Normal Productivity -- Tasks/Month/Person Effects -- Dimensionless
10/02/03
- ESD.36J SPM
39
Effect of Staff Experience 2
The effects are laws of nature ...

2 Effect of Quality of Prior Work 0 5 Years Experience 10
0 0 .5 1 Quality of Prior Work
What determines how the project moves up and down the curves?
Management Actions !

Progress
Work To Be Done
Quality
Staffing Requested
Added Work
Work Really Done
Known Rework
Undiscovered Rework
Obsoleted Work
Rework Discovery
Perceived Progress
10/02/03
- ESD.36J SPM
41
Major Themes to Date

Viewpoint on causes of dynamic behavior, and an introduction to the theory and tools to understand that behavior Insights into the dynamics of complex projects One model of the causes of those dynamics
- ESD.36J SPM Copyright 2003 James M. Lyneis 42
10/02/03
Are There Alternative Models?

More P & Q effects, etc. Variations on the basic rework cycle Multi-project and organizational models
10/02/03
- ESD.36J SPM
43
More Productivity and Quality Effects:

Model has 3 effects on P & Q :

Quality on quality Schedule pressure Experience Morale Organizational Size Changes Overtime Availability of supplier Sequence Other types of experience information and materials Skills match to needs ...
Copyright 2003 James M. Lyneis 44
What additional affects could be included?

10/02/03
- ESD.36J SPM
Other simplifications?
Task dependence/sequence is not represented explicitly -- with enough staff, could finish the project in a week Only one phase of work explicitly represented Suppliers are not represented Interactions with other projects are not represented These are treated endogenously or exogenously in more comprehensive models, and will be discussed in the remaining sessions.
When is the project finished?

In the current model, keep working until all work is completed correctly. In other situations, schedule may be more critical and therefore the project might:

reduce scope to meet schedule ship with errors
10/02/03
- ESD.36J SPM
46
Variations on the Basic Rework Cycle
Abdel-Hamid software development model Ford-Sterman product development model (see SD Bibliography on server)
10/02/03
- ESD.36J SPM
47
Broader Dynamics
Multi-project and organizational dynamics:
implementation of new tools and processes competition between concurrent and future projects for resources (We will discuss later in the term, but see Repenning and Sterman websites)
10/02/03
- ESD.36J SPM
48
Does the basic model apply to ...

Product development vs. one-of-akind construction projects? ... Hardware vs. software? Consumer products vs. military?
10/02/03
- ESD.36J SPM
49
Topics
10/02/03
- ESD.36J SPM
50
Brooks Law
"Adding manpower to a late software project makes it later." Brooks, Frederick P. Jr. The Mythical ManMonth. Reading, MA, Addison Wesley, 1995.
10/02/03
- ESD.36J SPM
51

l9 Dynamics 4

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

l9 Dynamics 4

Uploaded by

Copyright:

Available Formats

ESD.

36J System & Project Management

Dynamics of Project Performance

System Dynamics and Project Management Class Four (10/2/03)

Copyright 2003 James M. Lyneis

Copyright 2003 James M. Lyneis

Productivity Minimum Time to Perform a Task

Normal Quality Quality

Potential Work Rate Maximum Work Rate Feasible Work Rate

Quality on Quality Dynamics

Average Work Quality Maximum Time to Discover Rework

Time to Discover Rework Minimum Time to Discover Rework

Cumulative Effort Expended

<Project Finished Switch>

Copyright 2003 James M. Lyneis

Graph for Quality

Quality : Class3 Step1 Quality : Class2 Step2

Copyright 2003 James M. Lyneis

Graph for Effect of Prior Work Quality on Quality

Copyright 2003 James M. Lyneis

0.75 Fraction 75 Task

0.5 Fraction 50 Task

0.25 Fraction 25 Task

Fraction Fraction Task Task

Copyright 2003 James M. Lyneis

Graph for Work Done

Work Done : Class3 Step1 Work Done : Class2 Step2

Copyright 2003 James M. Lyneis

With short time to discover rework, quality is better ...

0.4 0 3 6 9 12 15 18 21 24 27 30 33 Time (Month) 36 39 42 45 48 51 54 57 60

Quality : Class3 Step1 Quality : Class3 Step1 Short TDRW

Copyright 2003 James M. Lyneis

because undiscovered rework is lower

0 0 3 6 9 12 15 18 21 24 27 30 33 Time (Month) 36 39 42 45 48 51 54 57 60 Task Task

Copyright 2003 James M. Lyneis

Less work needs to be done and the project finishes sooner

0 0 3 6 9 12 15 18 21 24 27 30 33 Time (Month) 36 39 42 45 48 51 54 57 60 Tasks Tasks

Copyright 2003 James M. Lyneis

Lesson: Its the undiscovered rework!

Copyright 2003 James M. Lyneis

Copyright 2003 James M. Lyneis

Staff and Schedule Dynamics

Copyright 2003 James M. Lyneis

Copyright 2003 James M. Lyneis

Copyright 2003 James M. Lyneis

Pressure team to work faster Work longer hours/overtime

Copyright 2003 James M. Lyneis

Anticipated Schedule Overrun

Scheduled Completion Date Normal Productivity Initial Scheduled Completion

Perceived Real Completion Date

<Time> Indicated Completion Date Based on Progress

Estimated Cost to Complete <Work to Do>

Average Productivity Copyright 2003 James M. Lyneis

Expected completion date:

Copyright 2003 James M. Lyneis

Effect of Schedule Pressure ...

up - Table for Effect of Schedule Pressur 2

Copyright 2003 James M. Lyneis

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Output = Input Table * Sensitivity + (1- Sensitivity)

Copyright 2003 James M. Lyneis

Schedule Pressure Simulations

Copyright 2003 James M. Lyneis