scholarly journals Distributed Stochastic Consensus Optimization With Momentum for Nonconvex Nonsmooth Problems

2021 ◽  
Vol 69 ◽  
pp. 4486-4501
Author(s):  
Zhiguo Wang ◽  
Jiawei Zhang ◽  
Tsung-Hui Chang ◽  
Jian Li ◽  
Zhi-Quan Luo
2000 ◽  
Vol 68 (1) ◽  
pp. 101-108 ◽  
Author(s):  
A. R. Hadjesfandiari ◽  
G. F. Dargush

A theory of boundary eigensolutions is presented for boundary value problems in engineering mechanics. While the theory is quite general, the presentation here is restricted to potential problems. Contrary to the traditional approach, the eigenproblem is formed by inserting the eigenparameter, along with a positive weight function, into the boundary condition. The resulting spectra are real and the eigenfunctions are mutually orthogonal on the boundary, thus providing a basis for solutions. The weight function permits effective treatment of nonsmooth problems associated with cracks, notches and mixed boundary conditions. Several ideas related to the convergence characteristics are also introduced. Furthermore, the connection is made to integral equation methods and variational methods. This paves the way toward the development of new computational formulations for finite element and boundary element methods. Two numerical examples are included to illustrate the applicability.


2013 ◽  
Vol 55 (2) ◽  
pp. 109-128 ◽  
Author(s):  
B. L. ROBERTSON ◽  
C. J. PRICE ◽  
M. REALE

AbstractA stochastic algorithm for bound-constrained global optimization is described. The method can be applied to objective functions that are nonsmooth or even discontinuous. The algorithm forms a partition on the search region using classification and regression trees (CART), which defines a region where the objective function is relatively low. Further points are drawn directly from the low region before a new partition is formed. Alternating between partition and sampling phases provides an effective method for nonsmooth global optimization. The sequence of iterates generated by the algorithm is shown to converge to an essential global minimizer with probability one under mild conditions. Nonprobabilistic results are also given when random sampling is replaced with points taken from the Halton sequence. Numerical results are presented for both smooth and nonsmooth problems and show that the method is effective and competitive in practice.


Author(s):  
Jan S. Hesthaven ◽  
Sigal Gottlieb ◽  
David Gottlieb

Sign in / Sign up

Export Citation Format

Share Document