Compete against a GA
Genetic algorithms, and metaheuristics in general, search for problem solutions by exploring in a stochastic way the search space. In general, this search is guided in order to improve efficacy and efficiency and not to waste computational resources in exploring regions of the search space with no promising solutions.
In the literature lots of work can be found in which metaheuristic techniques are applied with success to hard engineering problems. Usually, these problems are so hard that they are very difficult to solve without the computer assitance.
However, sometimes can be found surprising that these techniques, metaheuristics, using just a function to evaluate the quality of problem solutions, to be able to outperform human experts with their stochastic search. In this web page you have the oportunity of competing with a genetic algorithm.
In the left part of the applet below you can introduce binary strings with 10 digits. You can evaluate them using the Evaluate button. The result is shown in the corresponding label. You can evaluate as many strings as you want. After that you must push the Execute button in order for a simple genetic algorithm to solve the same problem using the same number of evaluations. The genetic algorithm will be executed 30 times and it will show in the right part of the applet statistics about the fitness values obtained in these 30 executions (maximum, minimum, and average). If you want to try it again you just have to push the Reset button. Who wins?
