摘要: In this work, we introduce a self-adaptive task farm for computational grids which is based on single-round scheduling algorithm called dynamic deal. principle, the deal approach employs skeletal forecasting information to automatically instrument and determine amount of work assigned each worker at execution time, allowing adapt effectively different load network conditions in grid. practice, it uses self-generated predictive values maps tasks onto nodes single-round. The effectiveness illustrated with biology parameter sweep non-dedicated departmental