Transcript

okay this is lecture 20 and this is the final lecture on determinants um and it's about the applications so we worked hard in the last two lectures to get a formula for the determinant and the properties of the determinant now to use the determinant and and always this determinant packs all this information into a into a single number and that number can give us formulas for all sorts of uh things that we've been calculating already without formulas now what was a inverse so so I'm be beginning with the formula for a inverse two 2x two formula we know right the 2x two formula for a inverse the inverse of a b c d inverse is one over the determinant times d a - b - c somehow I want to see what's going on for 3x3 and n byn and actually maybe you can see what's going on from this 2x two case so there's a formula for the inverse and what did I divide by the determinant so what I'm looking for is a Formula where it has one over the determinant and and you remember why that makes good sense because then that's perfect as long as the the determinant isn't zero and that's exactly when there is an inverse but now I have to ask can we recognize any of this stuff do you recognize what that number D is from the past from last from the last lecture my hint is think co-actors CU my formula is going to be my formula for the inverse is going to be one over the determinant times a matrix of co-actors so that you remember that D what's that the co-actor of remember co-actors if that's the one one co-actor because if I strike out row and column one I'm left with d and what's minus B okay which co-actor is that one Co minus B is the co-actor of C right if I strike out the C I'm left with a B there and why in the minus sign because this C was in a 2 one position and 2 + one is odd so there was a minus went into the co-actor and that's it okay I'll write down next what my formula is here's the big formula for the a for a inverse it's one over the determinant of a and then some Matrix and that Matrix is the Matrix of co-actors C only one thing I it turn you'll see I have to I transpose it so this is the Matrix of co-actors the what I'll just why don't we just call it the co-actor Matrix so the one one entry of of C is the co is the one one co-actor the thing that we get by throwing away row and column one it's the D and uh because of the transpose what I see up here is the co-actor of this guy down here right that's where the transpose came in what I see here this is the co-actor not of this one because I've transposed this is the co-actor of the B when I throw away the B the B row and the B column I'm left with c and then I have that minus sign again and of course the 22 entry is the co-actor of D and that's this a okay so there's the formula but we got to think why I mean it worked in this 2x two case but a lot of other formulas would have worked just as well we have to see why that's true in other words why is it so this is what I aim to to to uh find and and let's just sort of look to see what is that telling us that tells us that the formula that the expression for a inverse let's look at a 3X3 and I just write down a a b c d e f ghi and I'm looking for its inverse and what kind of a formula do I see there I mean what the determinant is a bunch of products of three factors right the determinant of this Matrix will involve a eii and bf time G and C * D * H and minus ceg and so on so things with three factors go in here things with how many F how many uh factors do things in the uh in The co-factor Matrix have what's a typical co-actor what's the co-actor of a the co-actor of a the one one entry up here in the inverse is I throw away the row and column containing a and I take the determinant of what's left that's the co-actor and that's EI minus FH products of two things I'm just making the observation the determinative of a involves products of n entries and the then the co-actor Matrix involves products of of n minus1 entries and like we never noticed any of this stuff when we were Computing the inverse by the ga Jordan method or whatever you remember how we did it we took the Matrix a we tucked the identity next to it we did elimination till a became the identity and then that the identity suddenly was a inverse well that was great numerically but we never knew what was going on basically and now we see what the formula is in terms of letters what's the algebra instead of the algorithm okay but I have to say why this is right right I still this is that's a pretty magic formula where does it come from well I'll just check it having having got it up there let me I'll I'll say how can we check what do I want to check I want to check that a times its inverse gives the identity so I I want to check that a * this thing a * this now I'm going to write in the inverse gives the identity so I check that a * C transpose let me bring the determinant up here determinant of a times the identity that's my job that that that's it that if this is true and it is then then I've correctly identified a inverse as C transpose divided by the determinant okay but why is this true why is that true let me let me put down what I'm doing here I have a here here's a here's A1 one I'm I'm doing this multiplication along to a1n and then down in this last row will be an a N1 along to a NN and I'm multiplying by The co-actor Matrix transposed so when I transpose it'll be C11 C12 down to c1n notice usually that one coming first would mean I'm in row one but I've transposed so that's those are the co-actors this First Column are the co-actors from Row one and then the last column would be the co-actors from row N and why should that come out to be anything good in fact why should it come out to be as good as this well you can tell me what the one one entry in the product is this is like you're seeing the main point if you just tell me one entry what do I get up there in the one one entry when I do this row of of ENT of this Row from a times this column of co-actors what what will I get there because we have seen this I mean we're right building exactly on what the last lecture reached uh so this is a11 * C11 a12 * C12 a1n * c1n what does that what does that sum up to that's the co that's the co-actor formula for the determinant that's the this co-factor formula which I wrote which I which I we got last time the determinant of a is if I use Row one let let I equal 1 then I have a11 * its co-actor a12 * its co-actor and so on and that gives me the determinant and it worked in this in this case uh this row times this thing is sure enough a D minus BC but this formula says it always works so up here in this in this position I'm getting determinant of a what about in the 2 two position row two * column 2 there what what is that that's just the co-actors that's just row2 * its co-actors so if of course I get the determinant again and in the last here this is the last row times its co-actors exactly you see we're realizing that the co-factor formula is just this sum of products so of course we think hey we've got a matrix multiplication there and we get determinant of a great but there's one more idea here right what else what have I not so I haven't got that formula completely proved yet because I've still got to do all the off diagonal stuff which I want to be zero right I just want this to be determinant of a times the identity and then I'm I'm a happy person so why should that be why should it be that one row times the co-actors from a different row which become a column because I transpose give Zer in other words the co-actor formula gives the determinant if the row and the and the co-actors you know if the if the entries of a and the co-actors are for the same row but for some reason if I take the co-actors from the fir the entries from the first row and the co-actors from the second row for some reason I automatically get zero and it's sort of like interesting to say why does that happen and can I just check that of course we know it happens uh in this case here are the numbers from Row one and here are the co-actors from row two right those are the numbers in row one and these are the co-actors from row two two cuz the co-actor of C is minus B and the co-actor of D is a and sure enough that row times this column gives please say it zero right okay so now how come how come can we even see it in this 2x two case why why well I mean I guess we in one way we certainly do see it because it's right here I mean do we just do it and we get zero but we want to think of some reason why the answer is zero some reason that we can use in the N byend case so let here here's my thinking we must be if we're getting the answer zero we suspect that what we're doing somehow we're taking the determinant of some Matrix that has two equal rows so I believe that if we multiply these by the co-actors from some other row we're taking the determinant yeah what Matrix are we taking the determinant of here it's this is it we're when we do that times this we're really taking can I put this in little letters down here I'm taking let me look at the Matrix a b a b let me call that Matrix a a s a screwed up okay all right so now that Matrix is certainly singular so if we find its determinant we're going to get zero but I claim that if we find its determinant by the co-actor rule go along the first row we would take a times the co-actor of a and what is the co so see how oh no I I'll let me go along the second row okay so let's see which if I take I know I've got a singular Matrix here and I believe that when I do this multiplication what I'm doing is using the co-actor formula for the determinant and I know I'm going to get zero let me try this again so the co-actor formula for the determinant says I should take a a Time its co-actor which is this B plus b * this time its co-actor which is this minus a okay that's what we're doing apart from a sign here oh yeah so you know there might be a minus multiplying everything so if I take this determinant it's a the determinant of this the determinant of a screwed up is a times its co-actor which is B plus the second guy times its co-factor which is minus a and of course I get the answer zero and this is exactly what's happening in that in that uh row times this wrong column okay that's the 2x two picture and it's just the same here that the reason I get a zero up in the is the reason I get a zero is that when I multiply the first row of a and the last row of the co-actor matrix it's as if I'm taking the determinant of this screwed up Matrix that has first and last rows identical the book pins this down more specific more carefully than I can do in the lecture I I hope you're seeing the point that this is an identity that that that it's a beautiful identity and it tells us what the inverse of the Matrix is so it gives us the inverse the formula for the inverse okay so that's the first goal of my lecture was to find this formula it's done okay and of course I can invert now uh I I I can I sort of like I can see what Happ I can answer questions like this suppose I have a matrix and let me move the one one entry what happens to the inverse just just think about that question suppose I have some Matrix I just write down some nice non- singular matrix it's got an inverse and then I move the one one entry a little bit I add one to it for example what happens to the inverse Matrix well this formula should tell me I have to look to see what happens to the determinant and what happens to all the co-actors and uh the picture it's it's all there it's all there we can really understand how the inverse changes when the Matrix changes okay now my second application is to let me put that over here is to ax equal B well of course the solution is a inverse B but now I have a formula for a inverse a inverse is one over the determinant times this C transpose time B I now know what a inverse is so now I just have to say what have I got here is there any way to to make this formula this this answer which is the one and only answer it's the very same answer we' got on the first day of the Class by elimination now I'm now I've got a formula for the answer can I play with it further to see what's going on and Kramer this Kramer's rule is exactly uh that a a way of looking at this formula okay so this is a formula for x here's my formula well of course the first thing I see from the formula is that the answer X always has that in the determinant I'm not surprised there's a division by the determinant but then I have to say a little more carefully what's going on up here here and let me tell you what Kramer's rule is let let me take X1 the first component so this is the first component of the answer then there'll be a second component and uh all the other components can I take just the first component of this formula well I certainly have determinant of a Down Under And what the heck is the first so what do I get in C transposed B what's the first entry of C transpose B that's what I have to ask myself well what's the first entry of C transpose B okay the this B is let me let me tell you what it is I I somehow I'm multiplying co-actors by the entries of B right in this product co-actors from The Matrix entries of B so anytime I'm multiplying co-actors by numbers I think I'm getting the determinant of something and let me call that something B1 so this is this is a matrix this is a this is a The Matrix whose determinant is coming out of that and we'll we'll see what it is X2 will be the determinant of some other Matrix B2 also divided by determinant of a so now I just Kramer just had a good idea he he realized what Matrix it was what these B1 and B2 and B3 and so on matrices were let me write them on the board underneath okay so what is this B1 this B1 is the Matrix that has B in its first column and otherwise the rest of it is a so it other wise it has the rest the The nus1 Columns of a it's it's it's the Matrix with it's it's just the Matrix a with column one replaced by the right hand side by the right hand side B because somehow when I take the determinant of this guy it's giving me this matrix multiplication well how could that be well how could so what's what's the determinant formula I'll use here I'll use co-actors of course and I might as well use co-actors down column one so when I use co-actors down column one I'm taking the first entry of B times what times the co-actor C11 do you see that when I when I use co-actors here I take the first entry here B1 let's call it times the co-actor there but what's that co-actor and by my little hand waving is meant to indicate that it's a matrix of one size smaller the co-actor and it's exactly C11 well that's just what we wanted the this first entry is C11 time B1 and then the next entry is whatever is C 21 * B2 and so on and sure enough if I look here when I'm when I do the co-actor expansion B2 is getting multiplied by the right thing and so on so there's Kramer's Rule and the book gives another kind of cute proof without uh without building so much on co-actors but here we've got co-actors so I thought I'd just give you this proof so what is B in general what is BJ this is the uh this is a with column J replaced by by B so that's that's the it's that DET it's the determinant of that Matrix that you divide by determinant of a to get XJ so X let me change this general formula XJ the jth one is the determinant of BJ divided by the determinant of a and now we've said what BJ is um well so Kramer found a rule and we could ask him okay great good work Kramer but is your rule any good in practice so he says well you couldn't ask a better rule than mine right because it's just all you have to do is find the determinant of a and these other determinants so I guess I just he just says well all you have to do is find n plus one determinants the the NBS and the A and actually uh I remember reading there was a book popular book that that kids interested in math read when I was a kid interested in math called mathematics for the million or something by a guy named Bell and it had a little page about linear algebra and it said uh so it explained elimination in a very complicated way I certainly didn't understand it and and it made it you know it sort of said well there is this formula for elimination but look at this great formula Kramer's rule so it certainly said Kramer's rule was the way to go go but actually Kramer's rule is a disastrous way to go because to compute these determinants uh takes like approximately forever so uh it's it's so I actually I now think of that book title as being mathematics for the millionaire because you'd have to be able to pay for uh a hopelessly long calculation where elimination of course produced the X's uh uh in an instant but having a formula allows you to with with letters you know allows you to do algebra instead of uh algorithms so there's some value in the Kramer's rule formula for x and in the explicit formula for for a inverse it's it's there're they're nice formulas but I just don't want you to use them that's that's what I comes to if you had to and mat lab would never never do it I mean it would use elimination okay now I'm ready for number three in today's list of amazing connections coming through the determinant and that number three is the fact that the determinant gives a volume okay so now so that's my final topic for the the among these it's this my number three application that the determinant is actually equals the volume of something can I use this little space to consider a special case and then I'll use the far board to think about the general rule so what am I going to prove I I I or or or claim I claim that the determinant of the Matrix is the volum volume of a box okay and you say which box fair enough okay so let's see I'm in shall we say we're in say 3x3 should we suppose let's let's say 3x3 so so we can really we're we're talking about boxes in three dimensions and 3x3 matrices and so all do just you could guess what the box is here's here's uh three dimensions okay now I take the first row of the Matrix a11 a22 a sorry a11 a12 a13 that that row is a vector it goes to some point that point will be and that edge going to it will be an edge of the box and that point will be a corner of the box so here here's 0 00 0 of course and here's the first row of the Matrix A1 1 a12 a13 so that's one edge of the Box another edge of the box is to the second row of the Matrix row two can I just call it there Row Two and a third row of the Box will be two the third row third edge of the Box will be given by Row three so so there's Row three that that that the coordinates what are the coordinates of that corner of the box a31 a32 a33 so I've got that edge of the box that edge of the box that edge of the box and that's all I need now I just finish out the box right I just make I make a sorry box the proper word of course is parallel a pipen but for obvious reasons uh I wrote box okay so uh okay so there's the there's the bottom of the box there are the four Edge sides of the Box there's the top of the box you right it's the box that has these three edges and then it's completed to a to a uh each you know each side is a is a parallelogram and it's that box whose volume is given by the determinant that's now it's possible that the determinant is negative so we have to just say what's going on in that case if the determinant is negative then the the volume we we should take the absolute value really so the volume if we if we think of volume is positive we should take the absolute value of the determinant but the the sign what does the sign of the determin it always must tell us something and somehow it it will tell us whether these three is a whether it's a right-handed box or a left-handed box if we if we reversed two of the edges we would go between a right-handed box and a left-handed box we wouldn't change the volume but we would change the the cyc the cyclic order so that I won't worry about that and so one special case is what a equal identity Matrix so let's take that special case a equal identity Matrix is the formula determinant of identity Matrix does that equal the volume of the Box well what is the Box what's the box if a is the identity Matrix then these three rows are the three coordinate vectors and the box is it's a cube it's the unit cube so if if a is the identity Matrix of course our formula is right well actually that proves property one that the volume has property one actually we could we could we could get this thing if we if we can show that the Box volume has the same three properties that Define the determinant then it must be the determinant so that's like the the the elegant way to prove this to prove this amazing fact that the determinant equals the volume first we'll check it for the identity Matrix that's fine the box is a cube and its volume is one and the determinant is one and and one agrees with one now let me take one let me go up one level to an orthogonal Matrix because I'd like to take this chance to bring in chapter the the previous chapter suppose I have an orthogonal Matrix what did that mean that I always called those things Q what was the point of suppose I have suppose instead of the identity Matrix I'm now going to take a equal Q an orthogonal Matrix what what was that what was Q then that was a matrix whose columns were orthonormal right those were its columns were unit vectors perpendicular unit vectors so what kind of a box have we got now what kind of a box comes from the rows or the columns I don't mind because the determinant is the determinant of the transpose so I'm never worried about that what kind of a box what shape box have we got if the Matrix is an orthogonal matrix it's another Cube It's a cube again how is it different from the identity Cube it's just rotated it's just the orthogonal Matrix Cube doesn't have to be the identity matrix it's just the unit cube but turned in space so sure enough it's a unit cube and its volume is one now is the determinant one what's the determinant of Q we believe that the determinant of Q better be one or minus one so that our formula checks out in that if we can't check it in these easy cases where we got a cube we're not going to get it in the general case so why is the determinant of Q equal 1 or minus one what do we know about Q what's the one Matrix statement of the properties of q a matrix with orthonormal columns has satisfies a certain equation what what is that it's if we have this orthogonal Matrix then the the fact this the the the way to say what it what it's properties are is this Q Prime Q transpose Q equals I right that's what those are the matrices that get the name Q the matrices that Q transposed Q is I okay now from that tell me why is the determinant one or minus one how do I out of this fact this may even be a homework problem it's there in the in the list of exercises in the book and let's just do it how do I get how do I discover that the determinant of Q is one or maybe minus one I take determinants on both sides everybody says so I won't I'll I won't I take determinants of both sides on the right hand side so I when I take determinants of both sides let me just do it take the determinant of take determinants the determinant of the identity is one what's the determinant of that product rule nine is paying off now the determinant of a product is the determinant of this guy maybe I'll put it y I'll use that symbol for determinant it's the determinant of that guy times the determinant of the other guy and then what's the determinant of Q transpose it's the same as the deter of Q rule 10 pays off so this is just this thing squared so that determinant squared is one and sure enough it's one or minus one great so in these special cases of Cubes we really do have determinant equals volume now can I just push that to non cubes let me push it first to rectangles rectangular boxes where I'm just multiplying the the edges are let me keep all the 90° angles because those are that that makes my life easy and just stretch the stretch the edges suppose I stretch that first Edge suppose this first Edge I double suppose I double that that first Edge keeping the other edges the same what happens to the volume it doubles right we know that the volume of a cube doubles in fact because we know that the new Cube would sit right on top of I mean the new the added Cube would sit right and would fit probably a geometer would say congruent or something would go right in in the other one we'd have two we have two identical cubes total volume is now two okay so if if I double an edge the volume doubles what happens to the determinant if I double uh the first row of a matrix what what's the effect on the determinant it also doubles right and that was rule number 3A remember rule 3 3 a was that if I I could if I had a factor in in row one t i could Factor it out so if if I have a factor two in that Row one I can Factor it out of the determinant it agrees with the the volume of the box has that factor to so so volume satisfies this property 3A and now I really am close but I but to get to the very end of this proof I have to get away from right angles I have to allow the possibility of uh other angles and or or what's saying the same thing I have to check that the volume also satisfies 3B so can I can I just so this is n of proof that so I'm determinant of a equals volume of box and where am I right now this volume has properties properties one no problem if the box is the cube everything is if if the box is the unit cube its volume is one property two was if I reverse two rows but that doesn't change the Box and it doesn't change the absolute value so no problem there property 3A was if I M you remember what 3A was so property one was about the identity Matrix property two was about a plus or minus sign that I don't care about property 3A was a factor T in a row but now I've got property 3B to deal with what was property 3B this is a great way to review these uh properties so that 3B the property 3B said Let's do let's do 2x two so said that if I had a + a prime b + B Prime C D that this equal what so this is property 3B this is the linearity in row one by itself so CD is staying the same and I can split this into a and a prime B Prime that's property 3B at least in the 2x two case and what am I I I wanted now to show show that the volume which two 2 by2 that means area has this has this property let me just emphasize that we have got we're getting this is a Formula then for the area of a parallelogram the area of this parallelogram can I just draw it okay here's the here's the parallelogram I have the row AB that's the first row that's the point ab and I tack on CD CD coming out of here and I complete the parallelogram so this is C so well I better make it look right it's really this one that has coordinates CD and this has coordinates well whatever the sum is and of course starting at 0 0 so we all know this is a plus c b plus d uh rather than I'm I'm pausing on that proof for a minute just to going back to our formula because I want you to see that unlike Kramer's rule that I wasn't that impressed by I'm very impressed by this formula for the area of a parallelogram and what's our formula what what's the area of that parallelogram if I had asked you that last year you would have said okay the area of a parallelogram is the base times the height right so you would have figured out what this base the how long that base was it's like the square root of A2 + b^2 and then you would have figured out how much is this height whatever it is it's horrible this I mean we got square roots and in that height there would be other revolting stuff but now what's the formula that we now know for the area it's the determinant of our little matrix it's just a D minus b c no square roots totally rememberable because it's exactly a formula that we've been studying the whole uh for three lectures okay that that that's um you know that's the most important point I'm making here is that if you know the coordinates of a box of the coiners then you have a great formula for the volume area or volume that doesn't involve any lengths or any angles or any Heights but just involves the coordinates that you've got and similarly what's the area of this triangle suppose I chop that off and say what about because you might often be interested in a triangle instead of a parallelogram what's the area of this triangle now there again everybody would have said the area of a triangle is half the base times the height and in some cases if you know the base that's that's and the height that's fine but here we what we know is the coordinates of the corners we know the vertices and so what's the area of that triangle if I know these if I know AB CD and 0 0 what's the what's the area it's just half so it's just half of this so this is this is AB a minus BC for the parallelogram and 1/2 of that 1 half of a D minus BC for the triangle so I I just I mean this is a totally trivial remark to say well divide by two but it's just that you more often see triangles and you feel you know the formula for the area but the good formula for the area is this one and I'm just going to J I'm just going to say one more thing about the area of a triangle it's just because it's it's you know it's so great to have a good formula for something what if our triangle did not start at 0 0 what if our triangle what if we had this what if we had so I'm coming back to triangles again but let me let me put this triangle somewhere it's I'm I'm saying with triangles I'm just in two Dimensions but I'm going to allow you to give me any three corners and in those six numbers must determine the area and what's the formula the area is going to be it's going to be there'll be that half of a parallelogram I mean basically this can't be completely new right we've got the area when we we know the area when this is 0 now we just want to lift our s slightly and get the area when all so let me write down what it what it comes out to be it turns out that if you do this X1 y1 and a one X2 Y 2 and a 1 X3 Y3 and a one that that works that that's determinant symbol of course it's just if I gave you that determinate to find you might subtract this Row from this it would kill that one subtract this row from this it would kill that one then you'd have a simple determinant to do with differences and it would it this little subtraction what I did re was equivalent to moving the triangle to start at the origin I did it fast because time is up and I didn't complete that proof of 3B I I'll leave the the book has a carefully drawn figure to show why that works but I hope you saw the main point is that for area and volume determinant gives a great formula okay and next lectures are about igen values so we're really into the big stuff thanks