I. S. Litvinchev Optimization , v.22, N1 , 01.1991 , p. 47-56, language: английский  Аннотация A decomposition-aggregation method is proposed for the solution of convex programming problems. In contrast to the known decomposition schemes, a special structure is not assumed for the original problem. It is shown that the problem of the adjustment of the aggregated problem in each iteration decomposes into independent subproblems of smaller dimension. If the original problem has block-, block-separable or block structure with coupling variables, then these subproblems are formulated in accordance with blocks.  Ключевые слова decomposition-aggregation method; block structure with coupling variable; convex programming; large-scale problems; computational methods; optimization