13 January 2020 in Cluster Computing
26 October 2020 in IEEE Access
With the development of artificial intelligence and the Internet of things, the prospects of cloud computing applications have become broader, and the number of users and cloud data centers (CDCs) has exploded. It is a challenge to realize efficient job scheduling and resource allocation of multiple users and data centers. However, the traditional scheduling model based on heuristic algorithm has some limitations in the complex and changeable cloud environment. In addition, many existing single-agent models rarely consider the multi-objective global optimization of the system. Therefore, this paper proposes a two-stage job scheduling and resource allocation framework that adopts multiple intelligent schedulers to solve the cooperative scheduling problem between job scheduling and resource allocation. A heterogeneous distributed deep learning (HDDL) model is used in the job scheduling stage to schedule multiple jobs to multiple cloud data centers. The deep Q-network (DQN) model is a resource scheduler to deploy virtual machine to physical servers for execution. Extensive numerical results show that both HDDL-baesd job scheduler and DQN-based resource allocator outperform the benchmark algorithm in terms of optimizing energy consumption and job delay. Furthermore, the proposed framework not only can achieve a global near-optimum by achieving local optimization at each stage but also has good scalability and low computation delay.
13 September 2023 in IEEE Access
At present, workflow scheduling in cloud computing environment is still a challenging optimization topic due to its NP-complete characteristics. In order to obtain better scheduling results, researchers are constantly coming up with new methods. In this study, we offer a hybrid metaheuristic for solving workflow scheduling in cloud to minimize the makespan of the workflow considering the heterogeneity of virtual resources. This hybrid approach combines the excellent optimization properties of Heterogeneous Earliest Finish Time (HEFT), Teaching–Learning-Based Optimization (TLBO), Opposition-Based Learning (OBL), and genetic manipulations, which is named Hybrid TLBO (HTLBO). Firstly, a HEFT-based method is proposed to produce the high-quality diverse initial population. Secondly, a Mixed OBL (MOBL) model is designed, in which the boundary search information and the population historical search information are systematically taken into account. Finally, an enhanced learner stage using genetic operations are added to effectively help the algorithm to jump out of the local optima. Rigorous experiments over various scientific workflows are conducted to validate HTLBO’s performance. The obtained results are compared to HEFT and some state-of-the-art hybrid metaheuristics in terms of average makespan, running time and non-parametric statistics. A significant improvement in schedule quality demonstrates that HTLBO can increase population diversity and achieve a good balance between scheduling effectiveness and efficiency.
27 June 2014 in Journal of Intelligent Manufacturing
4 December 2022 in Soft Computing
28 June 2018 in Applied Petrochemical Research
Currently, the manual method using hand-held infrared temperature measurement instruments for measuring temperatures on the external surfaces of ethylene cracking furnace tubes is highly subjective and is affected by a number of prominent issues, such as the high temperature working environments, which leads to low efficiency and poor measurement accuracy. Hence, an automatic temperature measurement system based on infrared light is designed and realized. In the system, a dual-phase drive synchronization method is proposed to rotate the thermodetector during horizontal movements, thus realizing automatic batch temperature measurements of the furnace tubes. Moreover, a temperature processing algorithm is developed to automatically identify furnace wall and tube surface temperatures, filter out abnormal temperatures and select only high-quality temperature measurements prior to calculating the final result. Real temperature measurement experiments demonstrated that the dual-phase drive temperature measurement system and temperature processing method are effective and efficient. Compared with the traditional manual way, temperatures obtained using the proposed system are more stable and accurate.