AI vs. AI — Virus Infection Battle with Minimax / Alpha-Beta Pruning

“Virus Infection Battle” pits two AI agents against each other on a 7×7 grid, each using Minimax with Alpha-Beta Pruning at a configurable search depth. The interesting part isn’t the game itself — it’s watching how search depth changes an agent’s behavior, and visualizing the number of nodes explored and iterations run by each algorithm as the depth increases.

Rules

Two players, Red and Blue, start with two tokens each on a 7×7 grid (Blue moves first). On a turn, a player can:

• Clone a token to an adjacent empty cell, or
• Jump a token two cells away, passing over other tokens.

Any opponent token adjacent to a cloned or jumped-to cell flips to the active player’s color. If a player has no legal move, they pass. The game ends when a player has no tokens left, both players pass consecutively, or a board state repeats — the winner is whoever has more tokens.

Why it was interesting

Implementing Minimax and Alpha-Beta Pruning from first principles — rather than using a game-AI library — made the classic trade-off between search depth and computation cost tangible: deeper search produces visibly smarter play, but the explored-node count grows fast enough that pruning stops being optional.

Stack: Java, custom Minimax/Alpha-Beta implementation. Code: github.com/VendenIX/JeuInfection Demo video: youtube.com/embed/oQxr1ursKIU