You are on page 1of 4

There are three planes, one that can carry 10 people, one that can carry 15 people, and

one that can carry 20 people. If 220 people are to y planes, what combinations work? First, as you did in the comments, create a system of equations: 10x + 15y + 20z = 220 x + y + z = 12 This can be made to be a matrix as follows:

10 1

15 1

20 1

220 12

The basic idea here is that you have the coecients on x,y,z rst, and the constant it is equal to last. It's also important that the same order of x,y,z used in each row is the same. Now, the row method works because of a simple idea: there are some things you can do to a system of equations that doesn't change the information: 1. multiplying an equation by something that isn't 0. For example, 5+2 = 7 also means that 10+4=14. The reason we restrict ourselves to nonzero numbers is that multiplying by zero destroys the information in an equation. This can be seen to be true, by noting that in the rst example, once we multiplied the equation by 2, we could undo this operation(just divide by two) So if I gave you the 10+4=14 equation and told you what operation I performed, you could tell me what equation I started with(thus we never lost any information!). This isn't true with the 0 + 0 =0 operation, though: if I multiplied by zero to get here, what equation did I start with? someone asks you this question, they're bullshitting you. Protip: if

There's simply no

way to gure that out; (this is why you can't divide by zero, btw. If division undoes multiplication, asking what number someone started with before they multiplied by zero is unfair). 2. Adding one equation to another. This one works because you can add any number to both sides of an equation and the equation is still true. if you have an equation relating two things to be equal, then they have the same value, so it's perfectly legit to add each side of one equation to each side of another(or even the same equation, but that would be the same thing as multiplying the equation by 2[Don't believe me? Try it out!]). And that's basically it; those are the things you can do to an equation without changing or removing information, so the only thing left to do is to gure out how this relates to the matrix we made. The rst thing to notice is that our matrix looks the same as our system of equations, except that we removed the variables, plus signs, and equal signs that were mucking things up. One important thing to note, though. In our matrix, there is one row for each equation, and once we are done with our matrix work, we'll want to be able to return to a system of equations, so we'll add the additional rule that we always keep the same number of rows and columns in our matrix. So the operations that we can do on this matrix that keep the information that we want are basically the same thing as the operations you can do to an equation: 1. replace a row with a constant multiple of itself

2.

replace a row with the sum of itself and a the result of multipling a

dierent row by some number 3. switch one row with another one. The rst operation corresponds exactly to the rst operation you can do to an equation. The second one is a little sneaky though: I said before that you can added two equations together, but whats with this multiplication jazz going on? What's happening here is that since we want to keep the same number of rows and columns in our matrix, we are allowing ourselves to do the rst and second operation at the same time. Multiplying an equation by a nonzero number doesn't change it, so we can then add the result of that equation to a dierent one and not lose any information(note that technically we allowed ourselves to multiply a row by zero, but since anytime we do this we are just adding the 0+0=0 equation, this won't hurt us this time. Since we didn't replace a row with the result of multiplication by zero, we're still good). The third operation is deceptively obvious: since the order of the equations never mattered to begin with, we can change the rows around as much as we like, and this operation can actually make some calculations less time consuming, so we'll keep it around just in case. So now on to our matrix:

10 1

15 1

20 1

220 12
Since we don't have

We are going to aim for the following: an equation for each variable of the form x=something y = something, or z = something. enough equations to get one answer, we won't be able to reach this goal, but we will end up with a simplied form of the system. Now since we intelligently decided to be consistent with the order of the variabes in each row, each column in the matrix corresponds to a dierent variable: the rst column corresponds to x, the second to y, the third to z, and the last column corresponds the =something(yay us!). This makes our goal easier to realize: if we transform our matrix into something that has only one 1 in each column, with all other numbers numbers in that column 0, and all the ones being in a dierent row, that will be the same thing as having one equation for each variable. But this owes some explanation: why is this the same thing as what we want? The answer is as follows: since we want x = something etc, we want the coecient on each variable to be 1. In other words, we want x = something, not 3x = something. That's why we have each column have a 1 in it. all the other numbers in a column being 0 is the same thing as having only one equation for each variable. In otherwords, we want x = something, not x = something and x = something else. The last condition is actually just to stop us from making a mistake. If all of our ones were in the same row, we would have to have the other rows be all 0, which would be the 0+0=0 equation!(or in our case, since we have three variables, 0+0+0=0). On that note, if you follow these rules and you end up with the 0 equation, don't panic. If you did everything right, it just means you had redundant information. Since our conditions ensure we have one and only one equation for each variable, this will force a redundant equation into the 0 equation form. Since we mandated the number of rows to stay the same, and

none of the operations delete or add information, there has to be an equation at the end where the redundant equation was, but it has to an equation that doesn't involve any variables, so it will have to be the 0 equation. Actually, it's also possible that you get a row like 0+0+0=something not zero. Why is this? This happens when the system you were working with was literally describing something impossible, like x + y = 2 and x + y = 3. This is actually another reason why matrices are useful; if you aren't sure if something can even happen in the rst place, just make a matrix for it and see if you get an impossible equation as a result. As a nal note, remember that in this system, we don't have enough equations to solve it completely, so we will have to settle with getting as close to our goal as we can. So now that we have a matrix, we know the rules, and we know what we're looking for, let's reduce that matrix! rst, let's switch the two rows, since that gives us a 1 in the top left corner for free. So we get:

1 10

1 15

1 20

12 220

Next, let's change that 10 on the bottom to a 0, so we only have on equation involving x in the end. Since there isn't a number such that 10 divided by it gives 0, we'll have to use the second rule here. So what number times 1 plus ten gives us 0? That's negative 10! So let's multiply the rst row by -10, and then add that result to the second row. Be careful to remember that the rst row isn't being changed at all; it's only being used to replace the second row. Only one row gets changed at a time. Here's the result:

1 0

1 5

1 10

12 100

So now we have an equation like x + something = something else, which isn't exactly, what we want. But it is as close as we can get in our case, due to the number of equations that we have. On the positive side, we have only one equation involving x, so hurray! So now we will want to change that 5 to a 1, so that we have a y = something equation. Dividing by 5 will do the trick:

1 0

1 1

1 2

12 20

And we only want one equation with y in it, so let's get rid of that top one in the second column. We'll want it to be 0, so we'll have to use the second rule here. since we want C*1+1 = 0, -1 is the number we'll want to multiply by. So we'll multiply the the second equation by -1, and then add the result to the rst equation:

1 0 0 1

1 2

8 20

Notice how getting the 0 in that rst column helped us out here: Since it's 0, any number we multiply it by will also be 0, so when we add the result of multiplication to a row, the 1 we worked so hard to get wasn't changed. This is we went in the order that we did: get a one, get the numbers above and below it to be 0, rinse and repeat with the other columns. Speaking of rinsing and

repeating though, we've gotten as close to our goal as we can. Thre's no way to change any of the numbers in the remaining columns to a 1 or a 0 without undoing work we've already done, so this is as close as we can get. So lets see if the system of equations is easier to comprehend now. When we started, we had the variables in the order of x, y, z, =something, so our matrix corresponds to the system: x -z = -8 y + 2z = 20 This system is denitely easier to understand than what we had originally. x and y are both expressed in terms of z and a constant, and they both only appear once, so when we think, what does x equal to? we don't have to worry about x being equal to something, and that value also satisfying a dierent equation altogether. This is a system that doesn't make my eyes bleed. And, we can now get rid of z altogether. We won't end up with a single answer, but we will get an equation between x and y that all solutions must satisfy. Notice that if the z in one equation is the negative of the z in the other equation, when we add the two equations, z will be turned into 0. So to do this, we will multiply the rst equation by 2: 2x - 2z = -16 And now we'll add this to the second equation: 2x + y = 4 we now have an easy way of checking if a combination is a solution: if it satises this equation, it's a solution. This an equation of two variables, so we'd have to have some additional information to go further... And we do: We have this simple fact: No one ies 2/3s of a plane when they don't have to, so x and y are whole numbers. Also, negative numbers of planes don't exist last time I wasn't in -1 world of mario, so x and y have to be positive as well. So x could be: 1, or 2 if y is 0. Next, we just plug these in to the equation to get what y is: 2, 0 And to get z, we go back to one of the original equations to get what z can be for these numbers: x+y+z=12 looks like the simplest; just plug in corresponding x and y values to get z: 1 + 2 + z = 12 z = 9 2 + 0 + z = 12 z = 10 So we have: x = 1, y = 2, z = 9 or x = 2, y = 0, z = 10 And now we're done! :]

You might also like