Please log in or register to post a reply.

No replies have been made yet.

- Upcoming Multiplatform Game Program...
- my new game 'squarecity' feedback w...
- Our first game - looking for feedbacks
- Network Emulation Tool
- Trouble with accessing GLSL array
- Fiction
- Game Programming Patterns: Bytecode
- Interactive WebGL Water Demo
- Skeletal Animation Tutorial with GP...
- Unreal Engine 4

Hi guys, currently I am working on software testing. In order to decide the test case to be used in a one testing process, it uses the historical data. The historical data is H(z_0, z_1…z_t;a_0,a_1…a_t). The value of zt is 1 if in the step t a failure detected and 0 if didn’t. and a_t is the value of the class of the test cases used in step t (1,2,3 or 4). The objective function is the minimum value of sum(z_i - r_i)\^2(i is from 0 to t).

r_i=(N - sum(z_j)*T_i ( j is from -1 to i-1 )

T_i is the probability of class i detect a failure…

I currently implement in C :

1. Generate initial population 100 binary string(0 0r 1) for z and 100 (1-4) for a randomly

2. Calculate corresponding z and a for fitness value f(H)

3. use the tournament selection method to generate new populations

4. Apply cross-over operator to the new populations and generate new populations

5. Apply mutation operator for each population.

I still need to modify, because the historical data must be used. In the second testing process the data from the first process should be used (so the initial population for the first element should not be random). Any idea for this implementation?

Thanks