Databases Reference
In-Depth Information
for each path, it is helpful to apply global optimization rather than local
search technique. It is because in local search technique, only local minima
for the branch function value may be found which might not be good enough
to traverse the desired branch. This problem can be solved by Genetic
Algorithms because it is a global optimization process.
In this chapter, partition testing is used to test each path. In partition
testing, a program's input domain is divided into a number of sub domains.
At the time of testing, the tester selects one or more elements from each
sub domain for analysis. The basic idea of partition testing is that each
path of the software belongs to certain sub domain, which consists of those
inputs which are necessary to traverse that path.
The domain is the set of all valid test sets. It is divided into sub domains
such that all members of a sub domain cause a particular branch to be
exercised. The domain notation may be based upon which branch (
true
or
false
) has been taken. The domain for the variables A and B of the
program in Fig. 10.2(a) is shown in Fig. 10.3. A character code shown
in Fig. 10.3 specifies the branch (here also path), e.g., TT (True True),
TF (True False), F (False), etc. In addition, the respective node is also
mentioned. In Fig. 10.3, the sub domain of node 5 is the dark grey area,
the sub domain of node 3 is the diagonal line, the sub domain of node 4 is
the light grey area whereas the sub domain of node 2 includes the light grey
area (node 2) plus the diagonal line (node 3). Domain testing tries to check
whether the border segments of the sub domains are correctly located by
(a)
(b)
Fig. 10.2.
(a) A sample program. (b) It's control flow graph.
