The linear system Ax = b is called homogeneous if b = 0. Otherwise, it is called inhomogeneous. These theorems can be applied to square systems of linear equations.
The linear system Ax = b is called homogeneous if b = 0. Otherwise, it is called inhomogeneous. These theorems can be applied to square systems of linear equations.
The linear system Ax = b is called homogeneous if b = 0. Otherwise, it is called inhomogeneous. These theorems can be applied to square systems of linear equations.
Theorems about homogeneous and inhomogeneous systems. Onthebasisofourworksofar,wecanformulateafewgeneralresultsaboutsquare systemsoflinearequations. Theyarethetheoremsmostfrequentlyreferredtointhe applications. Denition. ThelinearsystemAx =b iscalledhomogeneous ifb =0;otherwise,itis calledinhomogeneous. Theorem 1. Let A be an n n matrix. (20) |A| 0 A x =b has the unique solution, x = . = A 1 b (21) |A| 0 A x =0 has only the trivial solution, x = = 0. Notice that (21)is thespecialcaseof (20)whereb =0.Oftenitisstatedandusedinthe contrapositiveform: (21
) Ax =0 hasanon-zerosolution |A| = 0. (Thecontrapositiveofastatement P Q is not-Q not-P ; thetwostatementssay thesamething.) Theorem 2. Let A be an n n matrix. (22) |A| = 0 A x =0 has non-trivial (i.e., non-zero) solutions. (23) |A| = 0 A x =b usually has no solutions, but has solutions for some b. In (23),wecall thesystemconsistent ifithassolutions,inconsistent otherwise. Thisprobablyseemslikeamazeofsimilar-soundingandconfusingtheorems.Letsget another perspectiveonthese ideas byseeing howtheyapplyseparatelyto homogeneousand inhomogeneoussystems. Homogeneous systems: Ax =0 has non-trivial solutions |A| = 0. Inhomogeneous systems: Ax =b has the unique solution x =A 1 b, if |A| = 0. If |A| = 0, then Ax =b usually has no solutions, but does have solutions for some b. The statements (20)and (21)are proved,sincewe havea formula for thesolution,and it iseasytoseebymultiplyingA x =b byA 1 thatifx isasolution,itmustbeoftheform x =A 1 b. Weprove(22)justforthe3 3case,byinterpretingitgeometrically. Wewillgivea partialargument for (23), basedon bothalgebraand geometry. Proof of (22). WerepresentthethreerowsofA bytherowvectorsa, b, c andweletx = (x, y, z); thinkofalltheseasoriginvectors, i.e., place their tailsat theorigin.Then,consideringthe homogeneoussystemrst, (24) A x =0 isthesameasthesystem a x = 0, b x = 0, c x = 0. 1 2 HOMOGENEOUS AND INHOMOGENEOUS SYSTEMS Inotherwords,wearelookingforarowvectorx which is orthogonalto three givenvectors, namelythethreerowsofA.BySection1,wehave |A| =a b c =volumeofparallelepipedspannedbya, b, c. It followsthat if|A| = 0,the parallelepiped haszerovolume,and thereforetheoriginvectors a, b, c lieinaplane.Anynon-zerovectorx whichisorthogonaltothisplanewillthenbe orthogonaltoa, b, c,and thereforewill beasolution to thesystem (24).This proves (22): if|A| = 0,thenA x =0 hasanontrivialsolution. Partial proof of (23). WewritethesystemasA x =d, whered isthecolumnvector d = (d 1 , d 2 , d 3 ) T . Writingthisoutaswe did in (24), it becomes thesystem (25) a x =d 1 , b x =d 2 , c x =d 3 . If|A| =0,thethreeoriginvectorsa, b, c lieinaplane,whichmeanswecanwriteoneof them,sayc,asalinearcombinationofa andb: (26) c =ra +sb, r.s realnumbers. Thenifx isanyvector,itfollowsthat (27) c x =r(a x) + s(b x) . Nowifx isalsoasolution to (25),wesee from (25)and (27)that (28) d 3 =rd 1 +sd 2 ; thisshowsthatunlessthecomponentsofd satisfytherelation(28),therecannotbea solution to (25);thus in general therearenosolutions. Ifhowever,d doessatisfytherelation(28),thenthelastequationin(25)isaconse- quenceofthersttwoandcanbediscarded,andwegetasystemoftwoequationsinthree unknowns,whichwillingeneralhaveanon-zerosolution,unlesstheyrepresenttwoplanes whichareparallel. Singular matrices; computational diculties. Becausesomuchdependsonwhether|A| iszeroornot,thispropertyisgivenaname. WesaythesquarematrixA issingular if|A| = 0,andnonsingular orinvertible if|A| 0. = Indeed,we knowthatA 1 exists ifandonly if|A| = 0,whichexplainstheterm invertible;theuseofsingularwillbefamiliartoSherlockHolmesfans:itis the19thcenturyversionofpeculiarorthelate20thcenturywordspecial. EvenifA isnonsingular,thesolutionofAx =b islikelytorunintotroubleif|A| 0, orasonesays,A isalmost-singular.Namely,intheformulaforA 1 the|A| occursinthe denominator,sothatunlessthereissomesortofcompensationforthisinthenumerator, thesolutionsarelikelytobeverysensitivetosmallchangesinthecoecientsofA,i.e., to thecoecientsof theequations.Systems (ofanykind)whosesolutions behave thisway 3 HOMOGENEOUS AND INHOMOGENEOUS SYSTEMS aresaidtobeill-conditioned;itisdiculttosolvesuchsystemsnumericallyandspecial methodsmustbeused. Tosee the diculty geometrically, thinkofa 22systemAx =b asrepresentinga pair oflines;thesolutionisthepointinwhichtheyintersect. If|A| 0,butitsentriesare notsmall,then itstworowsmust bevectorswhicharealmost parallel (sincetheyspana parallelogramofsmallarea).Thetwolinesarethereforealmostparallel;theirintersection pointexists,butitspositionishighlysensitivetotheexactpositionsofthetwolines,i.e., tothevaluesofthecoecientsofthesystemofequations. MIT OpenCourseWare http://ocw.mit.edu 18.02SC Multivariable Calculus Fall 2010 For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.