Research on optimization of DAG task scheduling model based on heterogeneous multicore processor

Abstract

In order to better solve the problem of heterogeneous multi-core processors with tasks and complex relationships to scheduling problem based on the directed acyclic graph (DAG) model, through the task replication and merge operation, designed using a parallel task tree (PTT) model. The model chooses ant colony optimization (ACO) with natural parallelism to improve the task migration to complete the replication and merge of tasks, so as to give full play to the advantages of heterogeneous multi-core processors, and schedule tasks to corresponding nuclear execution. Then the comparison and analysis of the PTT model and the DAG model in the same parameters are carried out by using the speedup ratio and the task execution time as the index. The experimental results show that the PTT model has a high speed-up ratio, shortens the task execution time, strengthens the stability, and verifies the rationality of the design.