Using the randomizing algorithm, we can solve the job-shop scheduling problem MT20 in less than 100 iterations, with at most 10 backtracks per iteration. In contrast, MT10 is not satisfactorily solved.

The results of this program strongly confirm what we've already suggested: that there are many possible algorithms to solve a given type of problem, and that a combination of algorithms may get better results, on average, than each algorithm separately.