Derivative Free Analogues of the Levenberg-Marquardt and Gauss Algorithms for Nonlinear Least Squares Approximation.
Détermination rapide d'une solution de base initiale en programmation linéaire continue
Deux algorithmes de programmation mathématique
Deux problèmes d'affectation non linéaires
Differential evolution algorithm combined with chaotic pattern search
Differential evolution algorithm combined with chaotic pattern search(DE-CPS) for global optimization is introduced to improve the performance of simple DE algorithm. Pattern search algorithm using chaotic variables instead of random variables is used to accelerate the convergence of solving the objective value. Experiments on 6 benchmark problems, including morbid Rosenbrock function, show that the novel hybrid algorithm is effective for nonlinear optimization problems in high dimensional space....
Discrete, Nonlinear Approximation Problems in Polyhedral Norms.
Discrete, Nonlinear Approximation Problems in Polyhedral Norms. A Levenberg-Like Algorithm.
Distancias elipsoidales y puntos eficientes. Un método interactivo.
En este trabajo se estudian las propiedades que relacionan las distancias elipsoidales con la generación de puntos eficientes de un problema de programación multiobjetivo. Basándonos en estas propiedades, hemos construido un algoritmo interactivo convergente.
Domain optimization in -axisymmetric elliptic problems by dual finite element method
An axisymmetric second order elliptic problem with mixed boundary conditions is considered. The shape of the domain has to be found so as to minimize a cost functional, which is given in terms of the cogradient of the solution. A new dual finite element method is used for approximate solutions. The existence of an optimal domain is proven and a convergence analysis presented.
Dual method for solving a special problem of quadratic programming as a subproblem at nonlinear minimax approximation
The paper describes the dual method for solving a special problem of quadratic programming as a subproblem at nonlinear minimax approximation. Two cases are analyzed in detail, differring in linear dependence of gradients of the active functions. The complete algorithm of the dual method is presented and its finite step convergence is proved.
Dual method for solving a special problem of quadratic programming as a subproblem at linearly constrained nonlinear minimax approximation