Page 103 - 《软件学报》2026年第2期

582 软件学报 2026 年第 37 卷第 2 期

尽管 SAC-MWF 的有效性得到验证, 但其在目标场景中仍有一定改进空间. 其一, 虽然现有特征序列构建方
法在集成框架内能够实现预期目标, 但在长时预测任务上的效果有限. 因而需要寻找能够生成具备更强互补性的
特征序列的构建方法, 以进一步提高 SAC-MWF 的通用性. 其二, 在本文的设计中, 特征预测模型统一使用 MLP.
此举的目的是保证良好预测准确度的同时尽可能降低资源开销. 然而, 在综合考虑预测性能和资源开销时, MLP
不一定是最优模型. 因而可以进一步平衡集成模型的性能提升和运行时开销增加, 以构建出更具性价比的预测模型.

References
[1] Yang ZH, Wang XG, Li RT, Liu YL. HMM-CPM: A cloud instance resource prediction method tracing the workload trends via hidden
Markov model. Cluster Computing, 2024, 27(8): 11823–11838. [doi: 10.1007/s10586-024-04580-7]
[2] Gao JC, Wang HY, Shen HY. Task failure prediction in cloud data centers using deep learning. IEEE Trans. on Services Computing,
2022, 15(3): 1411–1422. [doi: 10.1109/TSC.2020.2993728]
[3] Rahmanian AA, Ghobaei-Arani M, Tofighy S. A learning automata-based ensemble resource usage prediction algorithm for cloud
computing environment. Future Generation Computer Systems, 2018, 79: 54–71. [doi: 10.1016/j.future.2017.09.049]
[4] Hameed A, Khoshkbarforoushha A, Ranjan R, Jayaraman PP, Kolodziej J, Balaji P, Zeadally S, Malluhi QM, Tziritas N, Vishnu A, Khan
SU, Zomaya A. A survey and taxonomy on energy efficient resource allocation techniques for cloud computing systems. Computing,
2016, 98(7): 751–774. [doi: 10.1007/s00607-014-0407-8]
[5] Baig SUR, Iqbal W, Berral JL, Carrera D. Adaptive sliding windows for improved estimation of data center resource utilization. Future
Generation Computer Systems, 2020, 104: 212–224. [doi: 10.1016/j.future.2019.10.026]
[6] Chen ZY, Hu J, Min GY. Learning-based resource allocation in cloud data center using advantage actor-critic. In: Proc. of the 2019 IEEE
Int’l Conf. on Communications. Shanghai: IEEE, 2019. 1–6. [doi: 10.1109/ICC.2019.8761309]
[7] Guo J, Chang ZH, Wang S, Ding HY, Feng YH, Mao L, Bao YG. Who limits the resource efficiency of my datacenter: An analysis of
Alibaba datacenter traces. In: Proc. of the 2019 Int’l Symp. on Quality of Service. Phoenix: ACM, 2019. 39. [doi: 10.1145/3326285.
3329074]
[8] Zhang QC, Yang LT, Yan Z, Chen ZK, Li P. An efficient deep learning model to predict cloud workload for industry informatics. IEEE
Trans. on Industrial Informatics, 2018, 14(7): 3170–3178. [doi: 10.1109/TII.2018.2808910]
[9] Hsu YF, Matsuda K, Matsuoka M. Self-aware workload forecasting in data center power prediction. In: Proc. of the 18th IEEE/ACM Int’l
Symp. on Cluster, Cloud and Grid Computing (CCGRID). Washington: IEEE, 2018. 321–330. [doi: 10.1109/CCGRID.2018.00047]
[10] Nguyen C, Klein C, Elmroth E. Multivariate LSTM-based location-aware workload prediction for edge data centers. In: Proc. of the 19th
IEEE/ACM Int’l Symp. on Cluster, Cloud and Grid Computing (CCGRID). Larnaca: IEEE, 2019. 341–350. [doi: 10.1109/CCGRID.2019.
00048]
[11] Tang XH, Liu QY, Dong YC, Han JZ, Zhang ZY. Fisher: An efficient container load prediction model with deep neural network in
clouds. In: Proc. of the 2018 IEEE Int’l Conf. on Parallel & Distributed Processing with Applications, Ubiquitous Computing &
Communications, Big Data & Cloud Computing, Social Computing & Networking, Sustainable Computing & Communications
(ISPA/IUCC/BDCloud/SocialCom/SustainCom). Melbourne: IEEE, 2018. 199–206. [doi: 10.1109/BDCloud.2018.00041]
[12] Chen L, Zhang WW, Ye HM. Accurate workload prediction for edge data centers: Savitzky-Golay filter, CNN and BiLSTM with
attention mechanism. Applied Intelligence, 2022, 52(11): 13027–13042. [doi: 10.1007/s10489-021-03110-x]
[13] Patel E, Kushwaha DS. A hybrid CNN-LSTM model for predicting server load in cloud computing. The Journal of Supercomputing,
2022, 78(8): 1–30. [doi: 10.1007/s11227-021-04234-0]
[14] Devi KL, Valli S. Time series-based workload prediction using the statistical hybrid model for the cloud environment. Computing, 2023,
105(2): 353–374. [doi: 10.1007/s00607-022-01129-7]
[15] Liu MH, Zeng AL, Chen MX, Xu ZJ, Lai QX, Ma LN, Xu Q. SCINet: Time series modeling and forecasting with sample convolution and
interaction. In: Proc. of the 36th Int’l Conf. on Neural Information Processing Systems. New Orleans: Curran Associates Inc., 2022. 421.
[16] Wang SY, Wu HX, Shi XM, Hu TG, Luo HK, Ma LT, Zhang JY, Zhou J. TimeMixer: Decomposable multiscale mixing for time series
forecasting. In: Proc. of the 12th Int’l Conf. on Learning Representations (ICLR 2024). Vienna: OpenReview.net, 2024.
[17] Zeng AL, Chen MX, Zhang L, Xu Q. Are Transformers effective for time series forecasting? In: Proc. of the 37th AAAI Conf. on
Artificial Intelligence. Washington: AAAI, 2023. 11121–11128. [doi: 10.1609/aaai.v37i9.26317]
[18] Haarnoja T, Zhou A, Abbeel P, Levine S. Soft actor-critic: Off-policy maximum entropy deep reinforcement learning with a stochastic
actor. In: Proc. of the 35th Int’l Conf. on Machine Learning. Stockholm: OpenReview.net, 2018. 1856–1865.
[19] Yang JQ, Liu CC, Shang YL, Cheng B, Mao ZX, Liu CH, Niu LS, Chen JL. A cost-aware auto-scaling approach using the workload
prediction in service clouds. Information Systems Frontiers, 2014, 16(1): 7–18. [doi: 10.1007/s10796-013-9459-0]
[20] Fang W, Lu ZH, Wu J, Cao ZY. RPPS: A novel resource prediction and provisioning scheme in cloud data center. In: Proc. of the 9th