Stochastic subspace correction methods and fault tolerance

2018/07/30 by Griebel, Michael, Oswald, Peter
#65F10 #65N22 #65N55 #65Y05 #68W20 #FOS: Mathematics #Numerical Analysis (math.NA)

paper · doi:10.48550/arxiv.1807.11315

Abstract

We present convergence results in expectation for stochastic subspace correction schemes and their accelerated versions to solve symmetric positive-definite variational problems, and discuss their potential for achieving fault tolerance in an unreliable compute network. We employ the standard overlapping domain decomposition algorithm for PDE discretizations to discuss the latter aspect.

Related

• Stochastic subspace correction in Hilbert space
• An Error-Resilient Redundant Subspace Correction Method
• Randomized Method of Subspace Corrections
• A Geometric Framework for Stochastic Iterations
• A Stochastic PCA and SVD Algorithm with an Exponential Convergence Rate