Search Based Software Engineering: Techniques, Taxonomy, Tutorial - Empirical Software Engineering and Verification - page 43

Table 7. Typification of the Software Cost Estimation Problem under the Proposed

Taxonomy

A. SOFTWARE ENGINEERING Perspective
1. Software Development Stage(s)
Software Planning
2. Software Development Model(s)
Waterfall Mo del
Spiral Model
Iterative and Incremental Development
Agile Development
3. Main Subject Descriptor(s)
D.2.9 Management
4. Main Implicit Subject Descriptor(s)
Cost Estimation

B. OPTIMISATION Perspective
1. Objective Space Dimensionality
Mono-objective
2. Instance Space Characterisation
Continuous
3. Constrained
No
4. Problem Linearity
Nonlinear
5. Base NPO Problem Type(s)
MISCELLANEOUS
6. Base NPO Problem(s)
BASIC OPTIMISATION PROBLEM

Step 1: The first step is to choose a representation of the problem and a fitness
function (see Section 3). The representation is important because it must be one
that can be readily understood; after all, you may find that you are examining
a great many candidate solutions that your algorithms will produce.
One should, of course, seek a representation that is suitable for optimisation.
A great deal has been written on this topic. For the first exploration of a new
SBSE application the primary concern is to ensure that a small change to your
representation represents a small change to the real world problem that your
representation denotes. This means that a small change in the representation
(which will be reflected by a move in a hill climb or a mutation in a genetic
algorithm) will cause the search to move from one solution to a 'near neighbour'