An approach to solving a problem of optimization with constraints is proposed. An algorithm based on a synthesis of such methods as the separation of variables, the dimension reduction, and the method of reducing the original problem to an auxiliary one. A number of applicability conditions for this algorithm and a convergence theorem are formulated.

Keywords: problems of nonlinear optimization, linearization method, linear constraints, reduces gradient method, linear convergence

• PostScript (in Russian) (506KB)
• PDF (in Russian) (133KB)



• PostScript. zip (in Russian) (270KB)
• PDF. zip (in Russian) (110KB)