Need help with genetic algorithm
Objective: There are five projects. Five tests for five different subjects. Every test has five questions. Users are tested and a table showing 0 or 1, i.e. correctly answered or not, is prepared on the basis of their responses. Users need to distribute projects on the basis of their responses. Each project could have five people. I don't know how to cross to get descendants, because basically we only have fixed data, and we can't change them and create new actors.
There was an idea that we just need to take as a gene all the answers to one test, and we already use the GA for each test separately. That is, for example, for each project, there is a certain estimate of what is needed for each test. And apply the GA to the test itself.
If the algorithm should result in the distribution of people by project, these distributions should be removed so as to obtain the best possible distribution by the given criteria. The results of the tests are needed to assess the allocations received in the GA process.