The rate of linear convergence of the Douglas-Rachford algorithm for subspaces is the cosine of the Friedrichs angle.
Heinz H. BauschkeJosé Yunier Bello CruzTran T. A. NghiaHung M. PhanXianfu WangPublished in: J. Approx. Theory (2014)
Keyphrases
- dynamic programming
- worst case
- cost function
- convergence rate
- objective function
- rapid convergence
- detection algorithm
- preprocessing
- optimal solution
- computational complexity
- search algorithm
- learning algorithm
- clustering method
- k means
- expectation maximization
- simulated annealing
- high accuracy
- computational cost
- np hard
- particle swarm optimization
- optimization algorithm
- probabilistic model
- high dimensional data
- euclidean distance
- parameter space
- recognition algorithm
- improved algorithm
- iterative algorithms
- faster convergence
- convergence property