Theory-inspired Parameter Control Benchmarks for Dynamic Algorithm Configuration.
André BiedenkappNguyen DangMartin S. KrejcaFrank HutterCarola DoerrPublished in: CoRR (2022)
Keyphrases
- learning algorithm
- dynamic programming
- optimization algorithm
- theoretical analysis
- data sets
- simulated annealing
- high accuracy
- np hard
- cost function
- detection algorithm
- k means
- recognition algorithm
- computational complexity
- computational cost
- probabilistic model
- improved algorithm
- control parameters
- clustering method
- ant colony optimization
- parameter tuning
- segmentation algorithm
- input data
- expectation maximization
- experimental evaluation
- significant improvement
- control system
- search space
- bayesian networks
- optimal solution
- classification algorithm
- parameter adjustment
- optimal parameters
- window size
- similarity measure
- preprocessing
- parameter space
- dynamic environments
- mathematical model
- graph cuts