Genetic Algorithm C# - Generic Implementation 

Thank you! I'm glad you enjoyed it :D

Next episode is out! :D

Yes sir! Planning to make the followup video this week or the next :) Oh, and if you want to go more in depth into the theory of genetic algorithms, I highly recommend checking out the playlist I linked in the description. It's from a channel called "The Coding Train", I thought he did an excellent job explaining it.

Thank you!

Kryzarel thank you!!!!

Oh crap, you're right, I've fixed the code in the repository and I'll mention it in the next video. Nice catch!

It should certainly stop faster, but are you sure it finds the correct answer?

You need only the DNA.cs and GeneticAlgorithm.cs classes. Create any type of Visual Studio project you wish and then just copy paste those files in there :)

For sure, use it for whatever you like!

will you make video how is it please

Probably Not. I'd prefer if you *suggest* videos and not *ask* for them. You would have to be a lot more specific in your request anyway. If this sounds rude at all I apologize, that is not my intention, I just like being straightforward.

Check out the link in the description for better understanding of how we're choosing parents. It's using a weighted random choice :)

@@Kryzarel I checked the link. I remember seeing this vid when it first game out and I'm just now coming back with a better understanding of programing. This part is still throwing me for a loop. I still don't quite understand why you're multiplying the rand.

I still don't quite get a few concepts from monster raising games and even if I did I'm still learning how to code so I 100% support this. His voice is pretty calm and his tutorials are really nice.

Just thought it would be better to let users have control over when and how they allocate the random objects. Maybe could have added an optional constructor that creates it if the user doesn't want to suppy one.

Thanks :)

We'll be messing with that script in the next video, I'll probably add the Unity project to the repository when that video goes live :)

Explained in this stack overflow answer: stackoverflow.com/questions/56692/random-weighted-choice It's an algorithm for choosing weighted random elements from a list. Link also in the description ;)

Thank you, I couldn't really understand why you were multiplying the fitnessSum with random.NextDouble() but after watching some videos I understand now.

And does it actually find the right answer or it just stops really fast with a random one? Could it be something to do with float/int conversions? The score variable is a float while targetString.Length is an int. Try casting targetString.Length to a float and see if it makes a difference. Also check if dna.Genes.Length == targetString.Length

It stops after a long time, finding a wrong answer. I checked everything you said. Still the same result. My only solution is to pow the score. for example, i get the score and i return 2^score as the final score. Then it works pretty well.

That seems quite weird. So the basic idea is to add 1 point to the score for each character that it gets right. If the target phrase is 10 characters long, then having a score of 10 means we have a phrase that has 10 characters right, meaning we got to the correct solution. To make it more generic I decided to divide that score by the length and get the percentage of "correctness" instead. If we had 5 characters right we'd get 5/10 = 0.5 so 50% correct, 10 characters right 10/10 = 1 so 100% correct. And after that there's yet another layer, where we use some exponents to make it so that having 6 characters correct is not just 10% better than 5 characters correct, but a bunch more. With this in mind I suggest you try to debug your code, maybe using breakpoints, and checking where in that process the calculations stop making sense. You can probably delete or comment out the last layer with the exponents to make the process simpler. If you don't mind sharing your code somehow, I'm also open to take a look at it.

I uploaded the code into a github repo if u wanna have a look: github.com/stanixdor/KryzarelGeneticAlgorithm

I took a look at the code and could not find anything obvious for what the issue might be. Other than this line ´if (score == 134)´, which I'm assuming is for debug purposes, everything seems to be alright. Strongly recommend you try to step through the code using breakpoints to try and figure out where the math stops making sense.

This is something different. I believe they both fall under the category of "Evolutionary Programming", but they are in fact two different algorithms. Sometimes they are used in tandem to improve each other, but still two different algorithms XD

It is based on this algorithm: stackoverflow.com/questions/56692/random-weighted-choice Should have had this as a link in the video, updating it now :P

Of course but I wouldn't write it like this. It's not written all that well.