On scheduling quality of precise resource measure sets which consist of tasks of circular type in GRID systems

Abstract

Abstract To serve parallel tasks of multiple users of Grid systems require the development of efficient methods for allocating processors and time resources. Optimal resource management is practically unrealizable due to the exponential complexity. In the situation, when the Grid-system is modeled by the resource quadrant and the user’s task is represented by the resource rectangle, heuristic polynomial algorithms are proposed with the quality assessment of scheduling by a Non-Euclidean heuristic measure. Using the example of arrays with an exact resource measure and induced by partridge linings, the adaptation of level polynomial algorithms is investigated, and a comparative analysis of the proposed algorithms is carried out.