home - Tetris AI

Tetris AI Challenge Specification

Develop a browser-based Tetris game in JavaScript that features an AI player powered by heuristic evaluation. The game must include:

Real-Time Gameplay: A Tetris board rendered on an HTML5 Canvas.
Heuristic-Based AI: The AI should simulate multiple moves (varying rotations and horizontal positions) and evaluate the resulting board states based on weighted metrics (aggregate height, complete lines, holes, and bumpiness).
Detailed Metrics Logging: A side panel next to the game must display real-time logs including:
- Training Time: Elapsed time since game start.
- AI Mode Status: Whether the AI is controlling the game.
- AI Move Count: Total number of moves taken by the AI.
- Latest Heuristic Score: The score of the most recent best-move simulation.
- Detailed Board Metrics: Aggregate height, number of complete lines, holes, and bumpiness for the best simulated move.

Speed Control:
- The game’s default (fast) mode should run with a fall speed of 500 ms per drop and an AI decision interval of 300 ms.
- Pressing the S key toggles a slow mode that multiplies both the fall speed and AI decision interval by a factor of 2 (making the gameplay slower for visualization).
- Pressing S again returns the game to the fast mode.
Control Modes: The AI should be the default controller (i.e. AI mode enabled at startup), with an option to toggle between human and AI control using the A key.
Clean Code: Use modular, well-commented code that cleanly separates game logic, AI logic, and UI rendering.

Languages/Technologies: JavaScript, HTML5, CSS.
Canvas Rendering: Utilize an HTML5 Canvas element to render the game board.
AI Logic:
- Implement functions to simulate potential moves, check for collisions, and lock pieces.
- Create evaluation functions to compute board metrics (aggregate height, complete lines, holes, bumpiness) and score each simulated move using weighted coefficients.
- The AI should then choose the move with the highest score and act accordingly.
Logging Panel: Render a side panel that updates in real time with the metrics listed above.
Speed Toggle: Implement a key handler for the S key to switch between fast and slow modes by adjusting the global timers for piece dropping and AI decision-making.

A single HTML file (or modular files) that implements the game and logs.
Clear inline documentation describing how each part of the code works.
A demonstration of the AI making decisions and the side panel updating with live metrics.
The ability to toggle between AI and human control (with the A key) and between fast/slow modes (with the S key).

Page updated

Google Sites

Report abuse