New Math Algorithm Solves Complex Games Instantly
This could make AI better at predicting outcomes in negotiations and security.
A deterministic polynomial-time algorithm computes an exact extensive-form proper equilibrium in a finite two-player zero-sum extensive-form game with perfect recall and rational data. Properness is understood in the agent normal form: an action at an information set is penalized relative to a better action while all other local behavioral choices, including the player’s own future actions, remain fixed.
The algorithm assigns each information set a geometric permutahedron of allowable action probabilities; a scaled doubly stochastic representation embeds these restrictions in a polynomial-size sequence-form polytope, and a saddle point then satisfies the local epsilon-properness inequalities. To obtain an exact equilibrium rather than a solution at an arbitrarily small numerical perturbation, it establishes sign stability of a parametric linear system, constructs a persistent feasible basis, and extracts the limits of behavioral ratios. The resulting probabilities are rational and have polynomial encoding length, including at information sets whose realization probabilities vanish.
- A new algorithm quickly finds unbeatable strategies for two-player games like chess or poker.
- It works for games where one player's win is the other's loss and players remember all moves.
- This could improve AI in cybersecurity, auctions, and negotiations by making decisions more reliable.
Why It Matters
Faster, exact game-solving could lead to smarter AI in security and finance, saving time and money.