An efficient algorithm for independent tasks scheduling in heterogeneous computing systems

Task scheduling is critical in heterogenous systems, especially with the huge number of tasks transmitted over grid. Since heuristics are proposing methods for solving heterogenous computing systems, several techniques were proposed for the scheduling on grid computing systems to get better execution time. In this thesis, a proposed new heuristic algorithm named Two Stages TasksTransfer (TSTT) algorithm introduced as an enhancement of state of TPB scheduling algorithm. Scheduling problem addressed Heterogeneous Computing Scheduling Problem (HCSP) mathematical model, where the independent tasks assigned to heterogenous processors with different characteristics. Twelve datasets with different heterogeneity level examined using different heuristic algorithms to compare the performance with the new algorithm. The proposed algorithm showed its efficiency in term of makespan, resource utilization and load balancing metrics for set of tasks