Author(s)

DanJu Lv^*, YueYun Yu, Yan Zhang

Affiliation(s)

College of Big Data and Intelligent Engineering, Southwest Forestry University, Kunming 650024, Yunnan, China.

Corresponding Author

DanJu Lv

ABSTRACT

Addressing the shift in AI-era optimization algorithms from offline static to integrated perception, this paper proposes a three-tiered pyramid teaching framework. By introducing computational graph perspectives to reshape foundational theory, empowering advanced training through dual-layer programming, and driving end-to-end differentiable optimization applications drives application practice, effectively resolving the challenges of fragmented knowledge, outdated tools, and single-dimensional assessment in traditional curricula. Practice demonstrates that this framework significantly enhances graduate students' dual-habitat cross-disciplinary problem-solving capabilities, bridging mathematical rigor and AI intuition.

KEYWORDS

Model-driven optimization; Integrated perception; Data-driven optimization; Bidirectional empowerment of AI and optimization; Prediction-decision integration

CITE THIS PAPER

DanJu Lv, YueYun Yu, Yan Zhang. Dual-engine drive of data + model on optimization theory for course for the AI era. World Journal of Educational Studies. 2026, 4(2): 17-21. DOI: https://doi.org/10.61784/wjes3132.

REFERENCES

[1] Watts J, Wyk F V, Khojandi M A. A dynamic deep reinforcement learning-bayesian framework for anomaly detection. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(12): 22884-22894.

[2] Wen J, Abeel T, De Weerdt M. Performance and interaction assessment of neural network architectures and bivariate smart predict-then-optimize. Machine Learning, 2025, 114(11).

[3] Qu K, Zhuang W. Scalable and dynamic cooperative perception: a data/model co-driven framework. IEEE Network, 2024(6): 38.

[4] Mandi J, Demirovic E, Stuckey P J, et al. Smart predict-and-optimize for hard combinatorial optimization problems. Proceedings of the AAAI Conference on Artificial Intelligence, 2020.

[5] Fan Z, Ghaddar B, Wang X, et al. Artificial intelligence for optimization: unleashing the potential of parameter generation, model formulation, and solution methods. European Journal of Operational Research, 2025,9(11).

[6] Balghiti O E, Grigas P, Elmachtoub A N, et al. Generalization bounds in the predict-then-optimize framework. Advances in Neural Information Processing Systems, 2020, 32: 3289-3299.

[7] Brahmi Z, Mezni H, Elmannai H, et al. Multi-modal vaas selection in smart mobility networks via spectral hyper-graph clustering and quantum-driven optimization. Concurrency & Computation: Practice & Experience, 2025, 37(23/24).

[8] Gasse M, Ferroni N, Chetelat D, et al. Exact combinatorial optimization with graph convolutional neural networks. Advances in Neural Information Processing Systems, 2020, 32: 15554-15566.

[9] Agrawal A, Amos B, Barratt S, et al. Differentiable convex optimization layers. Advances in Neural Information Processing Systems, 2019, 32: 9558-9570.

[10] Nandikolla V K, Ferman K, Barragan E, et al. Vision based obstacle detection and navigation of an autonomous vehicle. ASME International Mechanical Engineering Congress and Exposition. American Society of Mechanical Engineers, 2021, 85628: V07BT07A016.