Game theory; Adversarial bandit learning

The extended bandit learning game algorithm can search the best solution for the hybrid discrete-continuous strategy space. At each learning time, the player can quickly decide based on a finite discrete strategy pool, thereby improving the learning efficiency.