CIS 375 SOFTWARE ENGINEERING

University Of Michigan-Dearborn

Dr. Bruce Maxim, Instructor

 

Regression And Correlation.

Mean = (å xi)/n

Variance d = (å xi2)/n - (mean)2

Covariance m xy = (å xi yi)/n - (x)(y)

Correlation = r

Regression; y = mx + b

X

Y

X^2

X*Y

Y^2

1

1

1

1

1

3

2

9

6

4

4

4

16

16

16

6

4

36

24

16

8

5

64

40

25

9

7

81

63

49

11

8

121

88

64

14

9

190

126

81

S= 56

40

524

364

256

 

Regression Models

1. Linear Model

E = Co + S Cix

{Effort in Person Months} {Intercept} {Regression Coefficient}

2. Static Non-Linear Model

n

E = Co + S Cix xi2i

i=1

Often rewritten using size of code as a factor

E = a + b Sc m (x) n

{where a, b & c are constants, S = size, m = multiplier, and x = S Cix xi2i }

i=1

COCOMO Model

COnstructive COst MOdel

Contains 3 levels

1. Basic: Computes software development effort (and cost) as a function of program size, expressed in estimated lines of code.

2. Intermediate: Computes software development effort as a function of program size and a set of "cost drivers" that include subjective assessments of product, hardware, personnel, and project attributes.

3. Detailed: Includes all characteristics of the intermediate version with an assessment of the cost driver�s impact on each step (analysis, design, ect.) of the software engineering process.

Model: E = a Sb m(x)

 

BASIC

INTERMEDIATE

MODE

a

b

a

b

Organic

2.4

1.05

3.2

1.05

Semidetached

3.0

1.12

3.0

1.12

Embedded

3.6

1.20

2.8

1.20

in the basic COCOMO model x=1

in the intermediate model a questionnaire is used

Product Attributes

Computer Attributes

RELY

Reliability

TIME

Execution Time

DATA

Size of database

STOR

Amount of Storage

CPLX

Complexity of System

VIRT

Virtual Volatility

*

*

TURN

Response Time

Personnel

Project

ACAP

Analyst Capabilities

MODP

Modern Practices

AEXP

Analyst Experience

TOOL

Software Tools

PCAP

Programmer Capabilities

SCED

Development Schedule

PEXP

Programmer Experience

*

*

VEXP

Machine Experience

*

*

LEXP

Language Experience

*

*

Dynamic Estimation Models:

  1. Putnam Model (initially based on 200 military models).

(Rayliegh curve -> skewed, median & mean offset from one another).

Based on:
    1. Volume of work.
    2. Difficulty gradient (complexity).
    3. Project technology factor.
    4. Time constraints, delivery.

E = y(t ) = 0.3945 K;

{t represents optimal time} {K represents the area under the Rayliegh curve}.

D = K / t 2;

{difficulty measure based on t 2 and K}.

P = C2D-2/3;

{productivity measure}.

Ss = c K-1/5 t 4/3;

{ lines of code measure} {c represents the technology factor from survey}.

  1. Parr Model (variation of above):

Staff (t) = sech2 [(a t + c)/2].

  1. Jensen Model:

(less sensitive to compression)

Ss = Cte T K 1/2

{ Cte represents technology constraint with environmental adjustment}.
  1. COPMO (cooperative programming model):

E = E1 (S) + E2 (m);

{E represents total effort}.

{E1(S) represents the effort of one or people working independently

E1(S) = a + bs ,where a & b are empirically derived constants}.

{E2(m) represents the effort required to coordinate the development process

E2(m) = cm2 where m is the average number of team members}.

Make-Buy-Decision:

  1. Is a computer-based necessary (cost effective?):
  1. Can an existing software package be purchased for the task?

Decision Steps:

  1. Develop specifications.
  2. Estimate internal cost & delivery.
  3. Select 3 or 4 candidate packages.
  4. Select reasonable components.
  5. Build a cost-benefit comparison matrix, or Benchmarking (if from scratch you can�t do this).
  6. Evaluate each software package or component based on history with the product or vendor.
  7. Contact other users.

Now you�re in a good position to evaluate.

One way: build a decision tree.