EUVIMEDEUVIMED entdecken
Crossref

An efficient solution method for solving convex separable quadratic optimization problem

Li Shaoze, Wu Junhao, Lu Cheng, Deng Zhibin, Fang Shu-Cherng

Pacific Journal of Optimization · 2025 · 21 (4) · 677-689

DOI: 10.61208/pjo-2025-010 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Pacific Journal of Optimization, 2025. Verfasst von Li Shaoze und 4 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Convex separable quadratic optimization problems arise in numerous practical applications. These problems take the form: $\begin{align*}\min _{\boldsymbol{y} \in \mathbb{R}^n} f(\boldsymbol{y}) & =\boldsymbol{y}^T \Delta \boldsymbol{y}+\boldsymbol{\alpha}^T \boldsymbol{y} \\ \text { s.t. } g_{i}(\boldsymbol{y})& = \boldsymbol{y}^T \Theta_i \boldsymbol{y} +\boldsymbol{\beta}_i ^T \boldsymbol{y} +\sigma_i \leq 0,~ i=1,\dots,m, \label{SQPQC}\tag{SQPQC}\\\boldsymbol{y} & \in[\boldsymbol{l}, \boldsymbol{u}],\end{align*}$ where $\Delta$ and $\Theta_i$ are $n \times n$ diagonal matrices. Building on an iterative resolution scheme for the KKT system, we develop an efficient method for solving problem \ref{SQPQC}. We begin by applying the dual method for problem (SQPQC) with one single quadratic constraint, exploiting some distinct properties of this subclass, such as the differentiability of the dual function and the uniqueness of the optimal solution of the dual problem. Leveraging on the properties, we then develop a highly efficient algorithm for the general case of multiple quadratic constraints, and we show that the proposed approach leads to a dual coordinate ascent algorithm. Finally, we prove the convergence of the proposed algorithm and demonstrate, through extensive numerical experiments, that it outperforms the Gurobi solver, particularly for large-scale convex separable quadratic programming problems.

Zitieren

APA

Shaoze, L., Junhao, W., Cheng, L., Zhibin, D., & Shu-Cherng, F. (2025). An efficient solution method for solving convex separable quadratic optimization problem. Pacific Journal of Optimization, 21(4), 677-689. https://doi.org/10.61208/pjo-2025-010

Vancouver

Shaoze L, Junhao W, Cheng L, Zhibin D, Shu-Cherng F. An efficient solution method for solving convex separable quadratic optimization problem. Pacific Journal of Optimization. 2025;21(4):677-689. doi:10.61208/pjo-2025-010

BibTeX
@article{Shaoze2025,
  author = {Shaoze, Li and Junhao, Wu and Cheng, Lu and Zhibin, Deng and Shu-Cherng, Fang},
  title = {An efficient solution method for solving convex separable quadratic optimization problem},
  journal = {Pacific Journal of Optimization},
  year = {2025},
  volume = {21},
  number = {4},
  pages = {677--689},
  doi = {10.61208/pjo-2025-010}
}
Kostenlos starten →