A Priority-driven ACO Algorithm for DAG Task Scheduling in Cloud Environment

Abstract

Task scheduling in cloud environment is a key technical problem on how to allocate available cloud resources to cloud users. Usually, a task in cloud environment can be cut into a serials of subtasks, which have precedence and dependency relationships among themselves. We describe the problem by means of a DAG (Directed Acyclic Graph) model. Then, we propose a priority algorithm for DAG task scheduling, and a priority-driven ACO (Ant Colony Optimization) algorithm for DAG task scheduling on the basis of the DAG model. Finally, we compare these two algorithms with the greedy algorithm through simulation in the CloudSim platform. The simulation results show that the priority-driven ACO algorithm is effective to solve DAG task scheduling problem in cloud environment.