Computing Stackelberg Mixed Strategies and Applications to Security

Algorithms for computing game-theoretic solutions are now deployed in a number of real-world security domains, including air travel and harbors. These games are generally not zero-sum. Instead of Nash equilibrium, the preferred solution concept has been that of mixed Stackelberg strategies, which make sense when one player is able to commit to a distribution over actions before the other moves. I will review what we know about computing these strategies, both under standard game-theoretic representation schemes and under ones inspired by security games, as well as the relation of this solution concept to standard concepts including correlated equilibrium.