Random generation of string, towards goal result

[Bishop]

Hello,

Recently, I was experimenting a bit with genetic algorithms, and tried to generate simple string like "Hello World!". The algorithm works fine. But now I am wondering is there any mathematical or computer science topic, that covers solution for similar task without using genetic algorithms.

Like it is obvious that if it was no feedback that would be just brutual solution - random character generation until you hit the target, but lets say there is feedback that tells how close you are...

[WingedPanther]

Neural Networks are another feedback based system. With both Neural Networks and genetic algorithms, you generally are trying to avoid a brittle solution to a problem like generating a specific string. You want a little bit of variation to avoid predictability and have adaptability.

[Bishop]

Thing is that, I am not concerned about designing feedback part, what I am concerned about is doing same task without genetic algorithms.

[WingedPanther]

You can generally provide an evaluation function that will compare the output to the desired output and provide the neural net the required feedback.

[Predictor]

Hello,

Recently, I was experimenting a bit with genetic algorithms, and tried to generate simple string like "Hello World!". The algorithm works fine. But now I am wondering is there any mathematical or computer science topic, that covers solution for similar task without using genetic algorithms.

Like it is obvious that if it was no feedback that would be just brutual solution - random character generation until you hit the target, but lets say there is feedback that tells how close you are...

I am not sure exactly you are asking. There are other optimization systems which, like genetic algorithms, require feedback (evaluating their candidate solutions) such as simulated annealing, Nelder-Mead, etc.