The paper is concerned with computational research for large scale systems. The focus is on the hierarchical optimization methods that can be successfully applied to large scale optimization problems. A key issue is the possibility of solving several less dimension problems instead of one global high dimension task. Particular emphasis is laid on coarse granularity parallel implementation and its effectiveness. The paper discusses the usage of price coordination for real-life systems optimization. The results of numerical experiments performed for mean-variance portfolio selection using cluster of computers are presented and discussed.
展开▼
