CLOUD COMPUTING SCHEDULING IN DISTRIBUTED ENVIRONMENT USING DEEP LEARNING ALGORITHM

Task scheduling is a crucial element in determining the efficiency of cloud infrastructure and plays a key role in cloud computing. Task scheduling was described and graded as an NP-hard problem because of its difficulty. In addition, the most dynamic online work schedule also handles activities in a fluid environment, which makes it much more difficult that each part of cloud computing is balanced and satisfied. In this report, we suggest Deep reinforcement learning (DRL) that combines the advantages of the DRL algorithm with a strong neural network. Based on WorkflowSim innovations, the experiments take into consideration the variation in maps and load balancing in task scheduling relative to each task. Simulation are carried out to check the effectiveness of DRL optimization and learning skills and the results shows an improved performance by DRL than other methods.