Maze-Solving-AI-Agent

This repo basically has my journey to make an AI agent which solves mazes. I start our with the most basic agent - Simple reflex agent and move over to more complex agents like a model based reflex agent which has some percept history in it too. The code for all the agents can be found in the Agents folder where agent1 will be the most basic one and as we go higher they get better.

Different agents present so far:

1. Maze_Solver_Agent1 - It is a simple reflex agent with no random moves
2. Maze_Solver_Agent2 - It is a modal based agent with random moves and one level look back
3. Maze_Solver_Agent3 - It is a modal based agent with a percept history of last branch
4. Maze_Solver_Agent4 - It is a modal based agent which has a better memory as compared to agent3 and also stores the branches with previously visited sections of maze in it.

Simulations:

I have also written the code for visually simulating this agent and watching it run using pygame library. For more details on this open the 'Simulations' folder.

I will be adding more agents along the way if I come across some more ideas or find a better approach to it. Also your suggestion and ideas are always welcome you can open a PR or reach me using the details present on my ReadMe page.