List-Based Task Scheduling Algorithm for Distributed Computing System Using Artificial Intelligence

Abstract

Job scheduling in DAG (Directed Acyclic Graph) workflow have been a challenging task for the last couple of years. In DAG there is no miner so the time required to search task according to CPU is less. If an appropriate scheduling technique is not selected it may result in an increase in task execution time which may further negatively affect the energy consumption. Energy Prevention is one of the hottest issues in present era which is affecting the global environment. The problem of this research work is to propose a scheduling algorithm in such a manner that the consumption of energy for a DAG G (a, b), on the completion of all jobs is least. In this paper, an energy optimization model with the concept of task scheduling in cloud computing is proposed. List based HEFT (Heterogeneous Earliest Finish Time) algorithm is used to minimize the cost and energy consumption rate. On the basis of total execution time at every processor, the jobs are prioritized. On the basis of job priorities, neural network is trained. The neural network is used to classify the jobs on the basis of energy consumption. The jobs are assigning to the processor that consume less energy. At last the computed parameters such as energy consumption, SLR (Schedule length ratio) and CCR (Computation Cost Ratio) are measured.