<p>It is well known that the success of Lemke’s algorithm for solving a linear complementarity problem LCP(<i>q</i>,&#xa0;<i>M</i>) depends on the matrix class <i>M</i>. Many researchers investigated a large class of matrices for which Lemke’s algorithm computes a solution of the LCP(<i>q</i>,&#xa0;<i>M</i>). In this paper, we follow a different approach for the class of LCP, which is not solvable by Lemke’s algorithm. First, we construct an artificial LCP<InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="13226_2025_817_Article_IEq1.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="62" /> </InlineMediaObject> <EquationSource Format="TEX">\((\bar{q}_{1},\mathcal {M}_{1})\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mo stretchy="false">(</mo> <msub> <mover accent="true"> <mrow> <mi>q</mi> </mrow> <mrow> <mo stretchy="false">¯</mo> </mrow> </mover> <mn>1</mn> </msub> <mo>,</mo> <msub> <mi mathvariant="script">M</mi> <mn>1</mn> </msub> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> from LCP(<i>q</i>,&#xa0;<i>M</i>) by adding some artificial variables and extra constraints, and show that the matrix <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="13226_2025_817_Article_IEq2.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="28" /> </InlineMediaObject> <EquationSource Format="TEX">\(\mathcal {M}_{1}\)</EquationSource> <EquationSource Format="MATHML"><math> <msub> <mi mathvariant="script">M</mi> <mn>1</mn> </msub> </math></EquationSource> </InlineEquation> belongs to the class of semimonotone matrices. However, LCP<InlineEquation ID="IEq3"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="13226_2025_817_Article_IEq1.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="62" /> </InlineMediaObject> <EquationSource Format="TEX">\((\bar{q}_{1},\mathcal {M}_{1})\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mo stretchy="false">(</mo> <msub> <mover accent="true"> <mrow> <mi>q</mi> </mrow> <mrow> <mo stretchy="false">¯</mo> </mrow> </mover> <mn>1</mn> </msub> <mo>,</mo> <msub> <mi mathvariant="script">M</mi> <mn>1</mn> </msub> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> is not always solvable by Lemke’s algorithm. Then, we construct another artificial LCP<InlineEquation ID="IEq4"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="13226_2025_817_Article_IEq4.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="62" /> </InlineMediaObject> <EquationSource Format="TEX">\((\bar{q}_{2},\mathcal {M}_{2})\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mo stretchy="false">(</mo> <msub> <mover accent="true"> <mrow> <mi>q</mi> </mrow> <mrow> <mo stretchy="false">¯</mo> </mrow> </mover> <mn>2</mn> </msub> <mo>,</mo> <msub> <mi mathvariant="script">M</mi> <mn>2</mn> </msub> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> from LCP<InlineEquation ID="IEq5"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="13226_2025_817_Article_IEq1.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="62" /> </InlineMediaObject> <EquationSource Format="TEX">\((\bar{q}_{1},\mathcal {M}_{1})\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mo stretchy="false">(</mo> <msub> <mover accent="true"> <mrow> <mi>q</mi> </mrow> <mrow> <mo stretchy="false">¯</mo> </mrow> </mover> <mn>1</mn> </msub> <mo>,</mo> <msub> <mi mathvariant="script">M</mi> <mn>1</mn> </msub> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> by adding some more artificial variables and extra constraints that satisfy Eaves condition. We show that the resulting artificial LCP<InlineEquation ID="IEq6"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="13226_2025_817_Article_IEq4.gif" Format="GIF" Height="19" Rendition="HTML" Resolution="72" Type="Linedraw" Width="62" /> </InlineMediaObject> <EquationSource Format="TEX">\((\bar{q}_{2},\mathcal {M}_{2})\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mo stretchy="false">(</mo> <msub> <mover accent="true"> <mrow> <mi>q</mi> </mrow> <mrow> <mo stretchy="false">¯</mo> </mrow> </mover> <mn>2</mn> </msub> <mo>,</mo> <msub> <mi mathvariant="script">M</mi> <mn>2</mn> </msub> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation> is solvable by Lemke’s algorithm. Given an LCP(<i>q</i>,&#xa0;<i>M</i>), its solution can be obtained from the solution of the constructed artificial LCP(<InlineEquation ID="IEq7"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="13226_2025_817_Article_IEq7.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="51" /> </InlineMediaObject> <EquationSource Format="TEX">\(\bar{q}_{2},\mathcal {M}_{2}\)</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msub> <mover accent="true"> <mrow> <mi>q</mi> </mrow> <mrow> <mo stretchy="false">¯</mo> </mrow> </mover> <mn>2</mn> </msub> <mo>,</mo> <msub> <mi mathvariant="script">M</mi> <mn>2</mn> </msub> </mrow> </math></EquationSource> </InlineEquation>) with Eaves conditions. This approach leads to an innovative scheme for solving a large class of LCPs which are not solvable by Lemke’s algorithm. Further, we also provide convergence results. The results obtained here can be used for broader applications of Lemke’s algorithm.</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

On solving a larger subclass of linear complementarity problems by Lemke’s method

  • Sajal Ghosh,
  • Gambheer Singh,
  • Deepayan Sarkar,
  • S. K. Neogy

摘要

It is well known that the success of Lemke’s algorithm for solving a linear complementarity problem LCP(qM) depends on the matrix class M. Many researchers investigated a large class of matrices for which Lemke’s algorithm computes a solution of the LCP(qM). In this paper, we follow a different approach for the class of LCP, which is not solvable by Lemke’s algorithm. First, we construct an artificial LCP \((\bar{q}_{1},\mathcal {M}_{1})\) ( q ¯ 1 , M 1 ) from LCP(qM) by adding some artificial variables and extra constraints, and show that the matrix \(\mathcal {M}_{1}\) M 1 belongs to the class of semimonotone matrices. However, LCP \((\bar{q}_{1},\mathcal {M}_{1})\) ( q ¯ 1 , M 1 ) is not always solvable by Lemke’s algorithm. Then, we construct another artificial LCP \((\bar{q}_{2},\mathcal {M}_{2})\) ( q ¯ 2 , M 2 ) from LCP \((\bar{q}_{1},\mathcal {M}_{1})\) ( q ¯ 1 , M 1 ) by adding some more artificial variables and extra constraints that satisfy Eaves condition. We show that the resulting artificial LCP \((\bar{q}_{2},\mathcal {M}_{2})\) ( q ¯ 2 , M 2 ) is solvable by Lemke’s algorithm. Given an LCP(qM), its solution can be obtained from the solution of the constructed artificial LCP( \(\bar{q}_{2},\mathcal {M}_{2}\) q ¯ 2 , M 2 ) with Eaves conditions. This approach leads to an innovative scheme for solving a large class of LCPs which are not solvable by Lemke’s algorithm. Further, we also provide convergence results. The results obtained here can be used for broader applications of Lemke’s algorithm.