An Optimal Algorithm for Stochastic and Adversarial Bandits.
Julian ZimmertYevgeny SeldinPublished in: CoRR (2018)
Keyphrases
- dynamic programming
- optimal solution
- experimental evaluation
- computational complexity
- detection algorithm
- worst case
- monte carlo
- improved algorithm
- preprocessing
- ant colony optimization
- times faster
- np hard
- learning algorithm
- probabilistic model
- computational cost
- path planning
- cost function
- globally optimal
- exhaustive search
- recognition algorithm
- computationally efficient
- optimal strategy
- locally optimal
- weighting coefficients
- optimization algorithm
- theoretical analysis
- expectation maximization
- significant improvement
- tree structure
- convergence rate
- input data
- high accuracy
- multi objective
- optimal path
- search space
- multi armed bandit
- objective function