Initialising
Skip to content
RL Research

The Resourceful Rover

A 10×10 grid-world Mars colony where a Q-learning agent must harvest water ice under continuous resource depletion. The agent's state is deliberately projected down to (row, col) only — discarding water level and ice held — to study exactly what that simplification costs. Reproduces the original paper's training-curve result across 10 independent runs.

5,000 episodes × 10 runs68 pytest testsIn-browser demo

Q-learning agent reproduces the paper's reward curve: 10 runs × 5,000 episodes, 100-episode moving average climbs from a trough of −126.0 to a plateau of −106.4 (paper: ~−125 → ~−108). Try the same training loop live, in-browser.

2025RL Research
Interactive Demo

Train the agent, then watch it survive

The full environment and Q-learning loop below run in your browser — no server, no API calls. Click Train Agent to execute 5,000 episodes in about a second.

SYS.ONLINE·CH.03
T+20:11:45Z
Live RL Demo · Q-Learning in your browser
WATER
30
ICE
0/5
STEPS
0/150
REWARD
0
LAST
ROVERICEBASEOBSTACLE
SPEED200ms/step
Click Train Agent to run 5,000 Q-learning episodes in your browser and watch the learning curve emerge in real time.
// Honest note. The agent learns to navigate and deliver ice, but average reward plateaus below zero because the state representation is (row, col) only — it can't observe water level or ice held. That's the Curse of Dimensionality trade-off from the report, shown live: a richer state space would help, but at the cost of a much larger table and slower convergence.
Highlights
  • ε-greedy exploration decaying 0.4 → 0.01 (×0.9995/episode) over 5,000 episodes
  • Reward shaping: +10/ice unit delivered, +2 empty return, −1 move, −5 obstacle, −100 dry colony
  • Reproduces the paper's training curve to within ~2 reward units: trough −126.0 vs paper's ~−125, plateau −106.4 vs paper's ~−108
  • State deliberately projected to (row, col) only — the paper's central finding: this position-only state is why reward plateaus below zero instead of converging near zero
Case Study

How it works

01/04·Environment

10×10 Mars grid, depleting resources

The rover navigates a discretized 10×10 grid representing a Mars colony. Water reserves decay every step; ice patches replenish supply. Episode ends when reserves hit zero, or at a 150-step cap.

Next project
Build Your Ride