Finding the Smallest Eigenvalue by the Inverse Monte Carlo Method with Refinement

Abstract

Abstract. Finding the smallest eigenvalue of a given square matrix A of order n is computationally very intensive problem. The most popular method for this problem is the Inverse Power Method which uses LUdecomposition and forward and backward solving of the factored system at every iteration step. An alternative to this method is the Resolvent Monte Carlo method which uses representation of the resolvent matrix [I −qA] −m as a series and then performs Monte Carlo iterations (random walks) on the elements of the matrix. This leads to great savings in computations, but the method has many restrictions and a very slow convergence.In this paper we propose a method that includes fast Monte Carlo procedure for finding the inverse matrix, refinement procedure to improve approximation of the inverse if necessary, and Monte Carlo power iterations to compute the smallest eigenvalue. We provide not only theoretical estimations about accuracy and convergence but also results from numerical tests performed on a number of test matrices.

A Monte Carlo Method for Image Classification Using SVM

“…Monte Carlo methods are an established technique for solving problems where the unknown quantity can represented as the mathematical expectation of some random variable, so that by sampling this variable one can obtain estimates of the true value by averaging. In the domain of linear algebra the Monte Carlo methods have been used for solving linear systems, estimating eigenvalues or inverting matrices (see, e.g., (Alexandrov, 2005;Dimov, 1999)). In machine learning they have wide applicability since the problems at hand are inherently stochastic.…”

Quasi-Monte Carlo Methods for some Linear Algebra Problems. Convergence and Complexity

“…There has been renewed interest in MCMs in recent times, for example [18,11,12,10,7,13,1,2,3,9,8]. The primary reason for this is the efficiency of parallel MCMs in the presence of high communication costs.…”