Intelligent multi-machine scheduling on tasks with variable switch time

To find an optimal multi-machine schedule for a set of tasks with variable switch time, an optimal model was proposed, and a new intelligent algorithm with an adaptive structure was designed for less delay time and higher quality of service. If the server is under low average load, it used genetic algorithm and advanced greedy algorithm jointly to achieve a minimum task abandon ratio and a maximum task value, or else only advanced greedy algorithm was used to minimize delay time. The scheduler adopted runtime incremental parallel scheduling (RIPS) strategy, which combines the advantages of static and dynamic scheduling. The several simulation experiments show that the proposed schedule algorithm is valid and feasible.