A New Genetic Algorithm for Release-Time Aware Divisible-Load Scheduling

Divisible-load scheduling has become an increasingly important subject in the research of information technologies in recent years. It focuses on finding an efficient scheduling strategy for massive computing in parallel and distributed systems so that the make-span of the workload is minimized. Most existing scheduling models assume that all processors are idle at the beginning of workload assignment. However, in the real parallel and distributed environments, many processors may still in busy when a new workload arrived. Processors may have different waiting time from the busy state to the idle, that is, processors have different release time. This paper proposed a release-time aware divisible-load scheduling model with hybrid time constraints and designed an effective global optimization genetic algorithm to solve it. Finally, the experiment results show the efficiency and effectiveness of the proposed algorithm.