Skip to Main Content
Task scheduling has been proven to be NP-hard problem and we can usually approximate the best solutions with some classical algorithm, such as Heterogeneous Earliest Finish Time (HEFT), Genetic Algorithm. However, the huge types of scheduling problems and the small number of generally acknowledged methods mean that more methods are needed. In this paper, we propose a new method to schedule the execution of a group of dependent tasks for heterogeneous computing environments. The algorithm consists of two elements: An intelligent approach to assign the execution orders of tasks by task level, and an allocation algorithm based on chemical-reaction-inspired metaheuristic called Chemical Reaction Optimization (CRO) to map processors to tasks. The experiments show that the CRO-based algorithm performs consistently better than HEFT and Critical Path On a Processor (CPOP) without incurring much computational cost. Multiple runs of the algorithm can further improve the search result.