Game bots. How are they implemented?

Question

Bots

Hello. Bots have always fascinated me.
Recently, I have reviewed a lot of videos on the topics: "Neural Networks" and "Genetic Algorithm".
I really wanted to understand how bots work in races, for example.
After these videos and a couple of implementations of the genetic algorithm, I assumed (I had no idea at all before) that the bots work like this:
The neural network determines the state (how close we are to the wall, in which direction we are moving, the angle of rotation, etc.), the genetic algorithm learns from this data just get the best results. Most likely that such bots work simply on a neural network, and now I have 2 questions.
1) Am I right, or maybe my idea is at least close, or will it work at all?
2) I want to make a cs bot that will at least look like already implemented cs bots (against which we can play offline), are there any sources of those bots?

Answer 1

No.
It all depends a lot on the specific game.
There are no recipes.

Answer 2

genetic algorithm

Where did you get this from? Neural networks and genetics are completely different things. Even in biology, it's like a brick and a forget-me-not.
There are a lot of bots in the game, if each of them is based on neural networks - your game can only be launched on a supercomputer. So the developers make the illusion of the mind, not the mind. The more complex the logic of the bot, the more it loads the system and the more interesting it is as an opponent.
So:
1) no
2) first learn at least 1 programming language, then the basics of ai, etc. If you think that the source of something will give you an understanding of how it all works - no.
Igrostroy is interesting - start somewhere here:
https://unity3d.com/ru
https://www.cryengine.com/
https://www.unrealengine.com

Answer 3

Neural networks are not used in games, genetic algorithms are almost the same. Game AI is built on simple clear rules that are easy to set up and debug, and "smartness" is achieved through a variety of tricks that have nothing to do with intelligence. For example, an opponent who yells "I'm throwing a grenade" before throwing will appear smarter compared to the exact same but silent one.
In general, similar questions have already been asked, do not forget to use the search.

Answer 4

You don't understand materiel at all. Genetic Algorithm is an optimization technique. For example, neurons.
To get started, pull up the math and understand the basics of machine learning and neurons.
Now about bots.
If you mean game AI, then Daniil Basmanov is right .
If you mean by bots agents who play games, then everything is not so simple.
In simple games or games with a small space of moves, you can still get by with hard-coded logic, but in slightly more complex ones, you can no longer.
This is where neurons and Reinforcement Learning (reinforcement learning) come to the rescue. The most striking examples are Doom ai competition or a recent video of an OpenAI agent playing against a top player.