A PAC algorithm in relative precision for bandit problem with costly sampling.
Marie Billaud FriessArthur MachereyAnthony NouyClémentine PrieurPublished in: Math. Methods Oper. Res. (2022)
Keyphrases
- detection algorithm
- preprocessing
- dynamic programming
- experimental evaluation
- monte carlo
- learning algorithm
- sampling algorithm
- input data
- simulated annealing
- significant improvement
- recognition algorithm
- computational cost
- np hard
- times faster
- matching algorithm
- computational complexity
- linear programming
- theoretical analysis
- expectation maximization
- particle swarm optimization
- high accuracy
- objective function