Optimal pivoting strategy in LU factorization











up vote
0
down vote

favorite
1












I'm currently reading the book Numerical Linear Algebra by Lloyd N. Trefethen and David Bau, III, working my way through the required lectures for my Numerical Analysis class. The current subject is LU factorization, and while reading Lecture 21 about Pivoting (in Gaussian Elimination), I came across the statement that in practice, partial pivoting is equally as good as complete pivoting, while requiring inspection of a much smaller number of entries.



I'm pretty sure I understand correctly what partial and complete pivoting mean, namely with complete pivoting the entire $A_{k:m, k:m}$ submatrix is examined while with partial pivoting only the $A_{k:m, k}$ column is examined for an optimal pivot (the largest possible value, that is (right?)).



This seems very wrong to me and I'm surely missing some trivial argument for why this must be true. How can we be sure that part of the $k$th column of A always contains the biggest value in the entire submatrix? Am I right in assuming that the author meant to say that the increased gain from choosing the optimal value in the entire submatrix would be vastly overshadowed by the sheer amount of time needed to find it, and that only considering the first column of the submatrix yields good results?










share|cite|improve this question


























    up vote
    0
    down vote

    favorite
    1












    I'm currently reading the book Numerical Linear Algebra by Lloyd N. Trefethen and David Bau, III, working my way through the required lectures for my Numerical Analysis class. The current subject is LU factorization, and while reading Lecture 21 about Pivoting (in Gaussian Elimination), I came across the statement that in practice, partial pivoting is equally as good as complete pivoting, while requiring inspection of a much smaller number of entries.



    I'm pretty sure I understand correctly what partial and complete pivoting mean, namely with complete pivoting the entire $A_{k:m, k:m}$ submatrix is examined while with partial pivoting only the $A_{k:m, k}$ column is examined for an optimal pivot (the largest possible value, that is (right?)).



    This seems very wrong to me and I'm surely missing some trivial argument for why this must be true. How can we be sure that part of the $k$th column of A always contains the biggest value in the entire submatrix? Am I right in assuming that the author meant to say that the increased gain from choosing the optimal value in the entire submatrix would be vastly overshadowed by the sheer amount of time needed to find it, and that only considering the first column of the submatrix yields good results?










    share|cite|improve this question
























      up vote
      0
      down vote

      favorite
      1









      up vote
      0
      down vote

      favorite
      1






      1





      I'm currently reading the book Numerical Linear Algebra by Lloyd N. Trefethen and David Bau, III, working my way through the required lectures for my Numerical Analysis class. The current subject is LU factorization, and while reading Lecture 21 about Pivoting (in Gaussian Elimination), I came across the statement that in practice, partial pivoting is equally as good as complete pivoting, while requiring inspection of a much smaller number of entries.



      I'm pretty sure I understand correctly what partial and complete pivoting mean, namely with complete pivoting the entire $A_{k:m, k:m}$ submatrix is examined while with partial pivoting only the $A_{k:m, k}$ column is examined for an optimal pivot (the largest possible value, that is (right?)).



      This seems very wrong to me and I'm surely missing some trivial argument for why this must be true. How can we be sure that part of the $k$th column of A always contains the biggest value in the entire submatrix? Am I right in assuming that the author meant to say that the increased gain from choosing the optimal value in the entire submatrix would be vastly overshadowed by the sheer amount of time needed to find it, and that only considering the first column of the submatrix yields good results?










      share|cite|improve this question













      I'm currently reading the book Numerical Linear Algebra by Lloyd N. Trefethen and David Bau, III, working my way through the required lectures for my Numerical Analysis class. The current subject is LU factorization, and while reading Lecture 21 about Pivoting (in Gaussian Elimination), I came across the statement that in practice, partial pivoting is equally as good as complete pivoting, while requiring inspection of a much smaller number of entries.



      I'm pretty sure I understand correctly what partial and complete pivoting mean, namely with complete pivoting the entire $A_{k:m, k:m}$ submatrix is examined while with partial pivoting only the $A_{k:m, k}$ column is examined for an optimal pivot (the largest possible value, that is (right?)).



      This seems very wrong to me and I'm surely missing some trivial argument for why this must be true. How can we be sure that part of the $k$th column of A always contains the biggest value in the entire submatrix? Am I right in assuming that the author meant to say that the increased gain from choosing the optimal value in the entire submatrix would be vastly overshadowed by the sheer amount of time needed to find it, and that only considering the first column of the submatrix yields good results?







      numerical-methods numerical-linear-algebra gaussian-elimination lu-decomposition






      share|cite|improve this question













      share|cite|improve this question











      share|cite|improve this question




      share|cite|improve this question










      asked Nov 17 at 20:14









      Peiffap

      367




      367






















          1 Answer
          1






          active

          oldest

          votes

















          up vote
          1
          down vote



          accepted










          The authors do not claim that the partial and complete pivoting are equivalent and that the optimal pivot at each step of the elimination must be in the first column of the working sub-matrix.



          What they claim merely is that partial and complete pivoting strategies are equally good in the sense that the growth factor obtained by the partial pivoting strategy is not much worse than that obtained with the complete pivoting for vast majority of matrices arising in practice. There is no way how to prove this as there is no rigorous definition of what a practical matrix is and this claim is simply a result of decades of experience.



          There are well known examples showing that the partial pivoting can be much worse than the complete pivoting so this claim is not generally true.






          share|cite|improve this answer





















            Your Answer





            StackExchange.ifUsing("editor", function () {
            return StackExchange.using("mathjaxEditing", function () {
            StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
            StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
            });
            });
            }, "mathjax-editing");

            StackExchange.ready(function() {
            var channelOptions = {
            tags: "".split(" "),
            id: "69"
            };
            initTagRenderer("".split(" "), "".split(" "), channelOptions);

            StackExchange.using("externalEditor", function() {
            // Have to fire editor after snippets, if snippets enabled
            if (StackExchange.settings.snippets.snippetsEnabled) {
            StackExchange.using("snippets", function() {
            createEditor();
            });
            }
            else {
            createEditor();
            }
            });

            function createEditor() {
            StackExchange.prepareEditor({
            heartbeatType: 'answer',
            convertImagesToLinks: true,
            noModals: true,
            showLowRepImageUploadWarning: true,
            reputationToPostImages: 10,
            bindNavPrevention: true,
            postfix: "",
            imageUploader: {
            brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
            contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
            allowUrls: true
            },
            noCode: true, onDemand: true,
            discardSelector: ".discard-answer"
            ,immediatelyShowMarkdownHelp:true
            });


            }
            });














            draft saved

            draft discarded


















            StackExchange.ready(
            function () {
            StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3002763%2foptimal-pivoting-strategy-in-lu-factorization%23new-answer', 'question_page');
            }
            );

            Post as a guest















            Required, but never shown

























            1 Answer
            1






            active

            oldest

            votes








            1 Answer
            1






            active

            oldest

            votes









            active

            oldest

            votes






            active

            oldest

            votes








            up vote
            1
            down vote



            accepted










            The authors do not claim that the partial and complete pivoting are equivalent and that the optimal pivot at each step of the elimination must be in the first column of the working sub-matrix.



            What they claim merely is that partial and complete pivoting strategies are equally good in the sense that the growth factor obtained by the partial pivoting strategy is not much worse than that obtained with the complete pivoting for vast majority of matrices arising in practice. There is no way how to prove this as there is no rigorous definition of what a practical matrix is and this claim is simply a result of decades of experience.



            There are well known examples showing that the partial pivoting can be much worse than the complete pivoting so this claim is not generally true.






            share|cite|improve this answer

























              up vote
              1
              down vote



              accepted










              The authors do not claim that the partial and complete pivoting are equivalent and that the optimal pivot at each step of the elimination must be in the first column of the working sub-matrix.



              What they claim merely is that partial and complete pivoting strategies are equally good in the sense that the growth factor obtained by the partial pivoting strategy is not much worse than that obtained with the complete pivoting for vast majority of matrices arising in practice. There is no way how to prove this as there is no rigorous definition of what a practical matrix is and this claim is simply a result of decades of experience.



              There are well known examples showing that the partial pivoting can be much worse than the complete pivoting so this claim is not generally true.






              share|cite|improve this answer























                up vote
                1
                down vote



                accepted







                up vote
                1
                down vote



                accepted






                The authors do not claim that the partial and complete pivoting are equivalent and that the optimal pivot at each step of the elimination must be in the first column of the working sub-matrix.



                What they claim merely is that partial and complete pivoting strategies are equally good in the sense that the growth factor obtained by the partial pivoting strategy is not much worse than that obtained with the complete pivoting for vast majority of matrices arising in practice. There is no way how to prove this as there is no rigorous definition of what a practical matrix is and this claim is simply a result of decades of experience.



                There are well known examples showing that the partial pivoting can be much worse than the complete pivoting so this claim is not generally true.






                share|cite|improve this answer












                The authors do not claim that the partial and complete pivoting are equivalent and that the optimal pivot at each step of the elimination must be in the first column of the working sub-matrix.



                What they claim merely is that partial and complete pivoting strategies are equally good in the sense that the growth factor obtained by the partial pivoting strategy is not much worse than that obtained with the complete pivoting for vast majority of matrices arising in practice. There is no way how to prove this as there is no rigorous definition of what a practical matrix is and this claim is simply a result of decades of experience.



                There are well known examples showing that the partial pivoting can be much worse than the complete pivoting so this claim is not generally true.







                share|cite|improve this answer












                share|cite|improve this answer



                share|cite|improve this answer










                answered Nov 22 at 14:24









                Algebraic Pavel

                16.2k31839




                16.2k31839






























                    draft saved

                    draft discarded




















































                    Thanks for contributing an answer to Mathematics Stack Exchange!


                    • Please be sure to answer the question. Provide details and share your research!

                    But avoid



                    • Asking for help, clarification, or responding to other answers.

                    • Making statements based on opinion; back them up with references or personal experience.


                    Use MathJax to format equations. MathJax reference.


                    To learn more, see our tips on writing great answers.





                    Some of your past answers have not been well-received, and you're in danger of being blocked from answering.


                    Please pay close attention to the following guidance:


                    • Please be sure to answer the question. Provide details and share your research!

                    But avoid



                    • Asking for help, clarification, or responding to other answers.

                    • Making statements based on opinion; back them up with references or personal experience.


                    To learn more, see our tips on writing great answers.




                    draft saved


                    draft discarded














                    StackExchange.ready(
                    function () {
                    StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3002763%2foptimal-pivoting-strategy-in-lu-factorization%23new-answer', 'question_page');
                    }
                    );

                    Post as a guest















                    Required, but never shown





















































                    Required, but never shown














                    Required, but never shown












                    Required, but never shown







                    Required, but never shown

































                    Required, but never shown














                    Required, but never shown












                    Required, but never shown







                    Required, but never shown







                    Popular posts from this blog

                    Ellipse (mathématiques)

                    Quarter-circle Tiles

                    Mont Emei