Search

Task scheduling (TS) in cloud computing is a complex problem that involves balancing workload distribution, resource allocation, and power consumption. Existing methods often fail to optimize these objectives simultaneously and efficiently. This paper introduces a novel technique for scheduling independent tasks in cloud computing using multi-objective optimization and deep reinforcement learning (DRL). The proposed technique, DMOTS-DRL, combines Dueling deep Q-networks and dynamic prioritized experience replay to optimize two critical objectives: scheduling completion time (makespan) and power consumption. The performance of DMOTS-DRL is evaluated using CloudSim and compared with several state-of-the-art TS algorithms. The experimental results show that DMOTS-DRL outperforms the other algorithms in reducing makespan, power consumption, and other metrics, demonstrating its effectiveness and reliability for cloud computing services. Specifically, DMOTS-DRL achieves percentage improvements ranging from − 44.04 to − 0.19% in makespan, from − 0.26 to − 27.90% in power consumption, as well as better performance on other metrics such as energy consumption, degree of imbalance, resource utilization, and average waiting time. [ABSTRACT FROM AUTHOR]

Task scheduling is an attractive research topic in cloud computing nowadays. This process is very challenging and well known as NP-complete problem. Due to the dynamic and heterogeneous nature of user's request and provider's resource in cloud computing, the scheduling process still needs intelligent algorithms to achieve an efficient cloud resource allocation and to guarantee a good Quality of Service (QoS) for the users and their request classes. An important aspect for meeting these objectives is to design an effective task scheduling scheme which can not only satisfy users' varying priorities and QoS requirements, but also enhance providers' profit and system performances. In this paper, we introduce a new strategy to address the priority issue in both users' requests and providers' resources. We propose an efficient priority tasks scheduling called MCPTS, where the priority is adjusted according to four tasks' parameters including length, waiting time, deadline and burst time. MCPTS scheme consists of three sub-models such as tasks priority, task queueing priority and resources priority. A new hybrid multi-criteria decision-making (MCDM) method, namely ELECTRE III, and a meta-heuristic algorithm called differential evolution are proposed to evaluate and determine tasks' priorities. Further, we introduce a novel dynamic priority-queue algorithm based on queueing model. Furthermore, we adjust dynamically the resources priority based on tasks priority model in order to design an efficient and flexible relation between both resources and tasks classes. The proposed algorithm is validated through the CloudSim simulator. The experimental results indicate the superiority of MCPTS algorithm compared to other existing algorithms. Also, it shows the effectiveness of our algorithm in providing good system performance, satisfying users' priorities as well as QoS requirements, enhancing load balancing and improving resources utilization. [ABSTRACT FROM AUTHOR]