Parallel versions of the modified coordinate and gradient descent methods and their application to a class of global optimization problems (Q1395055)

The authors investigate the convergence of finite-difference local descent algorithms for the solution of global optimization problems with a multi-extremum objective function. The application of noise-tolerant local descent algorithms to the class of so-called \(\gamma\)-regular problems makes it possible to bypass minor extrema and thus to identify the global structure of the objective function. Experimental data presented in the article confirm the efficiency of parallel gradient and coordinate descent algorithms for the solution of some test problems.