An Extended Bandit Learning Game Approach

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. With the development of the learning time, the dynamic strategy pool can efficiently evolve to extend the whole hybrid discrete-continuous space, thereby avoiding missing the real best solution the hybrid discrete-continuous space.  Therefore, the proposed extended bandit learning game algorithm can achieve the quick search for hybrid discrete-continuous strategy spaces, and offers high applicability for the hybrid discrete-continuous resource optimization problem in unknown dynamic scenarios.