Skip to content
/ Wander Public

a Reinforcement learning Algorithm for solving Wander Problem

Notifications You must be signed in to change notification settings

Fsunroo/Wander

Repository files navigation

Wander

 

Wander

Github top language Github language count License

About   |   Puzzle Explanation   |   Requirements   |   Starting   |   Author


About

This project is a Reinforcement Learning Algorithm for Solving penniless pilgrim riddle puzzle

this simple app consist of two parts: 1- wanderEnv 2-WanderAgent

first one defines games enviroments, rules, moves and etc. while the latter is defining our Agents behaviour

0:Right 1:Down 2:Left 3:Up

the Agent will play the game more and more and each time learns a bit more of the game and completes his! Qtable, for more info pleas check Reinforcement learning

Puzzle Explanation

After months of travel, you’ve arrived at Duonia, home to the famous temple that’s the destination of your pilgrimage. The walk from the welcome center to the temple isn’t a long one … but there’s a problem. Can you outsmart the city’s imposed tax and make it to the temple without paying a fee?

Your tax total begins at 4. Walk east one block, add 2. Walk west a block, subtract 2. Walk south one block and your total doubles, but you can divide your total for every block walked northward.

How can you get to the temple with a zero balance? And without walking the same path twice? Grab a pencil and paper to solve this puzzle

Requirements

Before starting 🏁, you need to have numpy installed

Starting

# Clone this project
$ git clone https://github.com/fsunroo/wander

# Access
$ cd wander

# Install dependencies
$ pip3 install -r requirements.txt

# Run the project
$ python wanderAgent.py

Made with ❤️ by Mohammad Fhd

 

Back to top

About

a Reinforcement learning Algorithm for solving Wander Problem

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages