This is a re-implementation of a the NetLogo model Maze (ROOP, 2006).

This re-implementation makes use of the Q-Learning NetLogo Extension to implement the Q-Learning, which is done only with NetLogo native code in the original implementation.

This model implements a classic scenario used in Reinforcement Learning problem, the “Cliff Walking Problem”. Consider the gridworld shown below (SUTTON; BARTO, 2018). This is a standard undiscounted, episodic task, with start and goal states, and the usual actions causing movement up, down, right, and left. Reward is -1 on all transitions except those into the region marked “The Cliff.” Stepping into this region incurs a reward of -100 and sends the agent instantly back to the start (SUTTON; BARTO, 2018).

The problem is solved in this model using the Q-Learning algorithm. The algorithm is implemented with the support of the NetLogo Q-Learning Extension