Scholarlink

<p>This paper investigates the minimization problem of the sum of convex-concave quadratic fractional functions over a non-empty bounded convex quadratic feasible set, referred to as the SCQRP problem. Firstly, by fully considering the problem’s structure, it is transformed into an equivalent problem (EP). Subsequently, a convex relaxation programming problem (CRP) is formulated for the EP. Thirdly, we propose a branch-and-bound algorithm to solve the SCQRP globally. This algorithm generates a sequence of feasible solutions by solving a series of convex relaxation programming problems (CRP), wherein any cluster point of the feasible solution sequence is a global optimal solution. The paper theoretically proves the algorithm’s convergence and its worst-case computational complexity. Numerical experiments demonstrate the algorithm’s effectiveness and feasibility.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A outcome-space-based branch-and-bound algorithm for a class of sum of convex-concave quadratic ratio programs

  • 期刊论文

  • 发表日期：

    2026年3月11日

查看全文
  • Peng Hu,
  • Zhiyou Wu,
  • Hengyang Gu,
  • Shaojun Yin,
  • Xinyi Su

摘要

This paper investigates the minimization problem of the sum of convex-concave quadratic fractional functions over a non-empty bounded convex quadratic feasible set, referred to as the SCQRP problem. Firstly, by fully considering the problem’s structure, it is transformed into an equivalent problem (EP). Subsequently, a convex relaxation programming problem (CRP) is formulated for the EP. Thirdly, we propose a branch-and-bound algorithm to solve the SCQRP globally. This algorithm generates a sequence of feasible solutions by solving a series of convex relaxation programming problems (CRP), wherein any cluster point of the feasible solution sequence is a global optimal solution. The paper theoretically proves the algorithm’s convergence and its worst-case computational complexity. Numerical experiments demonstrate the algorithm’s effectiveness and feasibility.