Get documents about "lmpro ving Software Project Management System Dynamic Modeling
Document Sample
lmpro ving Software Project Management
Through System Dynamic Modeling
Creating schedules and tracking progress are two
activities software managers perform.
These activities become more complex with larger
projects.
. Wrong decisions implemented into a project could have
disastrous results, reducing the probability of success.
. Understanding the many system dynamics and feedback
loops associated with a project, especially a large
project, is a formidable task.
. There are plenty of Commercial Off The Shelf (COTS)
project management tools Available.
. Many perform basic features such as scheduling, budget
tracking, man-hour reporting, resource allocation, and
capital expenditure tracking.
. None of the tools available take into consideration how
system dynamic variables affect a project.
The following thesis is proposed:
Using the "iThinkntool available from High
Performance Systems, a new tool can be
developed to help a software manager plan, track,
and predict the outcome of a software project
better than the more traditional approaches by
incorporating system dynamic feedback loops
into the tool.
. "Staying the Course With the Project Control Panel"
by Susan Mann Beagley
4 Status Indicators: 4 Controllable Parameters:
n n
Resources
g Budget Work Seauence
wI
Percent Complete
nrvvl
t./ Quality
INCREMENTAL DESIGN APPROACH
Productivity
Defects - n
mP
Generafed .PmW
'i'
Schedule Pressure:
Rate of productivity of each level of Engineer. 10 - Don't Include; 1 Include - I
I inciudo Scneauie Press = I 1 J
0.0 ......... - g 1.0 T7
Eng 2 Productivity = 1-1 Communication Overhead:
00 =--a---= 2.0 v -
0 Don't Include; 1 Include -
Eng 3 Productivity = r j lnclucle Cornrn Overhead = I T ] I
0.0 p
=- 2.0 T7
Quality:
Quality will automatically generate defects that are generated by the engineers. These defects
are converted to hours to fix and filtered back into the model. Quality must be included
in order to include the reviews and/or rework %
10 - Don't Include: 1- Include
=
Include Qual~ty C'IiI Include Reiviews:
0.0 1.0
v. Quality must also be included for this to work.
When included the model will use 10% of the
Rework %: design and code phase for doing reviews.
l ~ u a l i must also be included for this to work.
t~ I
(percent of time allocated for rework. I -
0 Don't Include; 1 Include -
I Entered Rework = 1- 1 ] Include Review = (
1 1
errors per hr = tzzizl I
Increment #3
7
Number of Engineers Separated by the level 1
The scope of the work effort for each phase.
Level 3
expertise. Level 1 (less) --: (Highest) I
lnit Design 3 = 0
l l Eng 1 lncr 3 = I x a
O w . *. ' 2500 -p
0.0 5.0 'i7
Init Code 3 = rn Eng 2 lncr 3 = cxzl
0
- .
2500 0.0n
p 5.0 'i7-
Init Int 3 = 0 Eng 3 lncr 3 = m
-p
0.0 5.0 \7
- p p ~ ~
This is the level of messaging that will Task #3 Dependencies
-
0 No messages -
0 - -
1 None; 1 Design 2 Code; 3 Increment- I
-
1 Message Upon Completion of Task
-
2 1 + Msg after Code Phase Completes lncr 3 Dependent on 1 =
3 - 2 + Msg after Design Phase Completes
lncr 3 Msg Level = 1 lncr 3 Dependent on 2 = 1 7 1 I
Percent Complete Danqer Zone
I 1 : Danger Zone 2: Schedule % Comp
I % Complete: Page 1 Hours
SCHEDULE BUDGET
1: lncr 1 Scope 2: lncr 2 Scope 3: lncr 3 Scope 1 : Total Staff Load 2: Total Cost
Budget: Page 1 noun
PERCENT COMPLETE QUALITY
I : Danger Zone 2: Schedule X Comp 3: Porcelved X Comp 4: Actual X Comp 1: Total Allocated Rework 2: Rework & Defects 3: Total Rework 4: Total Defects
.
Y Complete: Page 1 Hours Quallty: Page 1 Houn
r schedule R Cornp 1 -I 1 L lncr 1 ScoDe I I 0.00 1 [ Rework R Defects / i
]
] Icliedule Pressure I 1 3.362 j
[ Actual %Cornp ] lncr 2 Scope ] 1 Work Rate Souqht ] [ Total Rework 1 7 1
[ ~ e r c e i v e d Camp
% 1-r
1 [ lncr 3 Scope I 1- I Work Rate 11 - I Total Defects 1] -
( Tolal Hours 1-1 I Exliauslion 11 13.43 1 Quality 1 -1 1
( Total Cost 1 I ~ a Tolerable Exh...
x I7 1
( Total Staff Load ] ( T I
Scenario #1 Scenario #2
1: Danger Zone 2: Schedule 96 Comp 3: Perceived % Comp 4: Actual % Cornp 1: Danger Zone 2: Schedule O/o Comp 3: Perceived % Cornp 4: Actual % Cornp
X Complete: Page 1 Houra I X Complete: Page 1 Hours
1: Total Allocated Rework 2: Rework & Defects 3: Total Rework 4: Total Defects I: Total Allocated Rework 2: Rework & Defects 3: Total Rework 4: Total Defects
Quality: Paqe 1 Hours Quality: Page 1 Hours
Scenario #I - Does Not Include Communication Overhead
Scenario #2 - Does Include Communication Overhead.
4%
JQ:*?. *
t
-- .-*
-
The validate approach.
. Identify a group of well qualified, reputable experts with good credentials
that are familiar with software engineering principles.
. Demo the tool with canned data to these experts, explaining the
implementation and characteristics of the tool.
. Give each evaluator the opportunity to make modifications to the canned
data.
. Give each evaluator the opportunity to run their own project supplied
data.
. Ask each evaluator to fill out and return an evaluation summary form.
. .
The first is to enhance the depth of the tool (For example, increase the
number of increments or better model the current feedback loops.)
. The second area is to improve the tool by adding breadth. (For example,
model the effects of various testing techniques on a project or adding
requirements analysis to the model.)
The third area is identifying other useful applications for the tool. (For
example, using it for training software managers.)
