Can be solved with Python 2.7 or Mathematica (10.2 or higher)
In this project you'll have to set up an environment (maze with walls 5x5), a robot and simulator functions. The robot has a fixed starting point and chooses its actions (right, left, up, down with certain probabilities) randomly. In the maze itself, some rewards are placed. In a second part, you'll have to implement a Q-learning algorithm and later experiment with different values of Q, alpha and epsilon.
Please find the whole exercise attached.
We'll need all codes of the solution.
10 фрілансерів(-а) подали заявки на цю роботу; середня заявка - €126
Hi, My last project here is very related to this one. Basically I'm an electronics engineer. I did the same thing with python. Come to chat for more discussion. Thank you
1. Studied Machine Learning, Neural Networks etc., 2. Sudied Markov decision process 3. Implemented Q-Learning before 4. Good at python, matlab