In this project you will have to create your own Tron AI snake.
### Getting started
You will need to create a public repository with the name `tron`. Next you need to create a file named `ai.js`. It must respect the instructions given.
### Controls
- `arrows` or `scroll` to move step by step
- `shift` will make it fast
- you can click anywhere on the progress bar to seek into the history
### Rules
- Your AI has to move every turn _(it can not stay still)_
- Every time the AI moves somewhere the AI leaves a color trail.
- the AI can only move to a blank tile.
- the AI can not move out of the map _(100 x 100)_
- the AI can only move to its `left`, `forward` or its `right`.
_(Moving `backward` is suicide as it would hit its own trail !)_
- If too much CPU power is required to decide where to go, the AI dies.
- If two AIs moved to the same spot, both of them die.
- **The AI has to survive as long as it can.**
### The game ends
- Once no players can make a move the player with the biggest score wins
Create your own Tron AI snake. You will be provided with a game engine, and your objective will be to create an AI player which will battle against other AI players to win the game.
The objective is for the AI player to stay alive as long as possible. The game ends when no player can move, or both players are dead.
### How to write your AI
The player with the highest score wins.
- AI players cannot stay still during a turn, they must make a move.
- Every time the AI player moves, it leaves a colorful trail.
- The AI player can only move to a blank tile, and cannot move outside of the 100 x 100 map.
- AI players can move `left`, `right` or `forward`. Moving backwards kills the AI player, because of the colorful trail.
- If too much CPU power is required to decide where to move, the AI player dies.
- If two AI players move to the same spot, both of them die.
### Controls
You can use the arrows or scroll to move step by step through the game. `shift` will make it fast. You can also click anywhere on the progress bar to seek through the history.
### How to write your AI
- Create `ai.js` at the root of your repository.
- Copy the contents of `random.js`, and paste it to `ai.js`.
- You may now edit the `update` function which is called each turn.
- Edit the `update` function which is called each turn.
> ⚠️ Do not rename the `update` function ⚠️ \
> as it's the function that the worker will try to run to test your AI.
> Do not rename the `update` function. It is the function that the worker will try to run to test your AI player.
### How to test your AI
**AI AT ROOT**
- You may test your ai locally. For that, create a folder and give it a name. Inside the folder you created, insert this file `index.html`.
- After that, create a simple web server by running the following command:
### Preparing to test your AI player
Let's say that you want to battle `ai.js` against `random.js`. You will need to run the game engine, and specify the local path or web path to both of those files.
```sh
$ &>/dev/null python3 -m http.server &
```
Now open your browser at the specified port. You'll use an appropriate command for your system:
For simplicity, we'll use a single directory to demonstrate. So place `index.html` and `random.js` in the same directory as `ai.js`.
- Linux: `xdg-open`
- macOS: `open`
- Windows: `start`
You'll need to create a simple web server. From the command line, `cd` into that directory, and execute the following command:
```sh
xdg-open 'http://localhost:8000'
$ python3 -m http.server
Serving HTTP on :: port 8000 (http://[::]:8000/)
```
- You can set a seed by adding the variable `seed` to the url params.
- You can add up to two AI's by adding the variable `ai` to the url params. The AI's will be separated by a `+`
Now open your browser at the specified port. For us, that will be where the `http.server` is running:
```
http://localhost:8000
```
- You can add a local file by specifying the relative path.
- You can add a online raw file by specifying the url to that file.
You can modify the path to load the players using a relative path:
```
http://localhost:8000/?ai=random.js+ai.js
```
A example of a url with local files using the default AI `ai.js` against the AI `hard.js` would be `http://localhost:8000/?ai=hard.js+ai.js&seed=2077349364`.
Each game has a `seed` which will spawn the players in different positions. When you modify the path, it will add a `seed` to the query path. You can change the `seed` to spawn the players elsewhere, or remove it for a random one to be generated.
A example of a url with online files would be `http:?/localhost:8000/?ai=https://((DOMAIN))/git/root/public/raw/branch/master/subjects/tron/ai/hard.js+ai.js`.
You can also specify a web path to some `.js` file like so:
Note: You can test a local AI against and online one.
Open the inspector of the browser and **disable the cache**.
- Open the inspector of the browser used and **disable the cache**
- let's change the update function so that your AI only goes forward.
### Getting started with improving your AI player
Let's change the `update` function so that your AI only goes forward.
Replace this line just before the `return` of the update function:
@ -99,17 +87,11 @@ return pickRandom(available);
return coordsInBound.filter(isFree)[0];
```
- save the file, push the changes and re-run the game in the browser.
If the cache was correctly disabled,
you have changed your AI behaviour from a random pick of available moves
to only going forward.
Save the file, and re-run the game. If the cache was correctly disabled, you will have changed the behavior of your AI player from random movements, to only making forward moves.
- To understand better the way of controlling your AI,
read the comments inside the AI file and do a lot of testing.
To better understand controlling your AI player, read the comments inside `ai.js`, and do a lot of testing.
- When peer-corrected, you AI will be competing against other AIs.
Be aware that there will be the possibility for the peer-correcter
to use his or her own AI.
During the audit, you will be competing against other AI players. Be aware, that there is a possibility for the auditor to use their own AI.
davhojt
2 years ago
committed by