Parallel Implementation of Task Scheduling using Ant Colony Optimization

Efficient scheduling of tasks for an application is critical for achieving high performance in heterogeneous computing environment. The task scheduling has been shown to be NP complete in general case and also in several restricted cases. Because of its key importance on performance, the task scheduling problem has been studied and various heuristics are proposed in literature. This paper presents a novel framework for task scheduling problem based on Ant colony optimization (ACO). The inherent parallelism of this heuristics is exploited to be implemented effectively on multicore processors. The performance of the algorithm is demonstrated by the time taken for producing effective schedules for random task graphs.