<p>DR (Diminishing Return)-submodular functions, which capture diminishing marginal benefits, are widely applicable, especially when incorporating a linear regularization term to effectively address individual costs and sparsity in the objective maximization. This paper is to maximize a regularized DR-submodular function <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40305_2025_633_Article_IEq1.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="174" /> </InlineMediaObject> <EquationSource Format="TEX">\(F = H + L: [0,1]^n \mapsto \mathbb R\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>F</mi> <mo>=</mo> <mi>H</mi> <mo>+</mo> <mi>L</mi> <mo>:</mo> <msup> <mrow> <mo stretchy="false">[</mo> <mn>0</mn> <mo>,</mo> <mn>1</mn> <mo stretchy="false">]</mo> </mrow> <mi>n</mi> </msup> <mo>↦</mo> <mi mathvariant="double-struck">R</mi> </mrow> </math></EquationSource> </InlineEquation> over a compact, convex and solvable set <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40305_2025_633_Article_IEq2.gif" Format="GIF" Height="15" Rendition="HTML" Resolution="72" Type="Linedraw" Width="46" /> </InlineMediaObject> <EquationSource Format="TEX">\(\mathcal Y \ni \textbf{0}\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi mathvariant="script">Y</mi> <mo>∋</mo> <mn mathvariant="bold">0</mn> </mrow> </math></EquationSource> </InlineEquation>, where the function <i>H</i> is non-negative, differentiable and DR-submodular, while function <i>L</i> is linear. When <i>H</i> is monotone, we propose an approximation algorithm achieving a bi-factor type approximation ratio of <InlineEquation ID="IEq3"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40305_2025_633_Article_IEq3.gif" Format="GIF" Height="23" Rendition="HTML" Resolution="72" Type="Linedraw" Width="70" /> </InlineMediaObject> <EquationSource Format="TEX">\(\left( 1-\frac{1}{\textrm{e}},1\right) \)</EquationSource> <EquationSource Format="MATHML"><math> <mfenced close=")" open="("> <mn>1</mn> <mo>-</mo> <mfrac> <mn>1</mn> <mtext>e</mtext> </mfrac> <mo>,</mo> <mn>1</mn> </mfenced> </math></EquationSource> </InlineEquation>. When <i>H</i> is non-monotone, we present two corresponding approximation algorithms for both case that <i>L</i> is non-positive and unrestricted, with approximation ratios as <InlineEquation ID="IEq4"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40305_2025_633_Article_IEq4.gif" Format="GIF" Height="23" Rendition="HTML" Resolution="72" Type="Linedraw" Width="64" /> </InlineMediaObject> <EquationSource Format="TEX">\(\left( \frac{1}{4}, \log 2\right) \)</EquationSource> <EquationSource Format="MATHML"><math> <mfenced close=")" open="("> <mfrac> <mn>1</mn> <mn>4</mn> </mfrac> <mo>,</mo> <mo>log</mo> <mn>2</mn> </mfenced> </math></EquationSource> </InlineEquation> and <InlineEquation ID="IEq5"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="40305_2025_633_Article_IEq5.gif" Format="GIF" Height="23" Rendition="HTML" Resolution="72" Type="Linedraw" Width="42" /> </InlineMediaObject> <EquationSource Format="TEX">\(\left( \frac{1}{4}, \frac{1}{2}\right) \)</EquationSource> <EquationSource Format="MATHML"><math> <mfenced close=")" open="("> <mfrac> <mn>1</mn> <mn>4</mn> </mfrac> <mo>,</mo> <mfrac> <mn>1</mn> <mn>2</mn> </mfrac> </mfenced> </math></EquationSource> </InlineEquation>, respectively. Finally, we present numerical experiments that support the approximate guarantee conclusions of this paper and validate the effectiveness of the proposed algorithms compared to existing methods.</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

Maximization of DR-Submodular Regularization Under Convex Constraints: A Study of Bi-Factor Approximation Algorithms

  • Zhi-Yuan Dong,
  • Yi-Han Wang,
  • Yang Zhou

摘要

DR (Diminishing Return)-submodular functions, which capture diminishing marginal benefits, are widely applicable, especially when incorporating a linear regularization term to effectively address individual costs and sparsity in the objective maximization. This paper is to maximize a regularized DR-submodular function \(F = H + L: [0,1]^n \mapsto \mathbb R\) F = H + L : [ 0 , 1 ] n R over a compact, convex and solvable set \(\mathcal Y \ni \textbf{0}\) Y 0 , where the function H is non-negative, differentiable and DR-submodular, while function L is linear. When H is monotone, we propose an approximation algorithm achieving a bi-factor type approximation ratio of \(\left( 1-\frac{1}{\textrm{e}},1\right) \) 1 - 1 e , 1 . When H is non-monotone, we present two corresponding approximation algorithms for both case that L is non-positive and unrestricted, with approximation ratios as \(\left( \frac{1}{4}, \log 2\right) \) 1 4 , log 2 and \(\left( \frac{1}{4}, \frac{1}{2}\right) \) 1 4 , 1 2 , respectively. Finally, we present numerical experiments that support the approximate guarantee conclusions of this paper and validate the effectiveness of the proposed algorithms compared to existing methods.