Reinforcement Learning Lab

Q-Learning

Alpha (LR): 0.3Gamma: 0.90

Exploration

Epsilon: 0.20Show Policy Arrows

Speed

Steps/tick: 5

Stats

Episodes

Avg Reward

0.0

Grid

6x6

States

Reinforcement Learning Lab – Agents, Rewards & Policies

Reinforcement learning is a branch of machine learning where an agent learns to maximize cumulative reward by exploring an environment and receiving feedback on its actions. Unlike supervised learning, there are no labeled examples—the agent discovers what works through trial and error, gradually improving its policy. This lab lets you define the environment, reward structure, and learning parameters, then watch the agent's behavior evolve from random exploration to purposeful, near-optimal play.

What you can do in this simulation

Choose a grid-world or custom environment and define cell-level rewards and penalties
Watch the agent explore randomly during early training and exploit learned paths later
Visualize the Q-table or policy grid updating after each episode
Adjust exploration rate (epsilon) and discount factor (gamma) and observe convergence speed
Plot cumulative reward per episode to diagnose learning stability

Concepts covered

reinforcement learning · Q-learning · reward function · policy · exploration vs exploitation · Markov decision process

Related simulations

Browse all Technology simulations →

Free to use in your browser — no signup required. Found a bug or have an idea to make it better? Tell us.

Please wait

Loading technology simulations

The requested content is still loading.

Finished experimenting?

Save one completion to this browser's signed guest ledger, or directly to your account when signed in.