1

Let $U_1, U_2 $ be sub spaces in $R_4[x]$, such that:

$$U_1 = Sp\{x^3+2x^2+3x+6, 4x^3-x^2+3x+6, 5x^3+x^2+6x+12\}$$ $$U_2 = Sp\{x^3-x^2+x+1,2x^3-x^2+4x+5\}$$

Find $U_1 \cap U_2$


My idea (i need help how to proceed): compare linear combination of both spanning sets, get the general solution for the homogeneous system.

In the end im not sure, its becoming too complicated technically.

Let $a_1,a_2,a_3,a_4,a_5 \in \Bbb R$:

$$a_1(x^3+2x^2+3x+6)+a_2(4x^3-x^2+3x+6)+a_3(5x^3+x^2+6x+12) = a_4(x^3-x^2+x+1)+a_5(2x^3-x^2+4x+5)$$

$$\Downarrow$$

$$a_1(x^3+2x^2+3x+6)+a_2(4x^3-x^2+3x+6)+a_3(5x^3+x^2+6x+12) - a_4(x^3-x^2+x+1) - a_5(2x^3-x^2+4x+5) = 0$$

$$\Downarrow$$

$$\begin{bmatrix}1&4&5&-1&-2&|0 \\ 2&-1&1&1&1&|0 \\ 3&3&6&-1&-4&|0 \\ 6&6&12&-1&-5&|0\end{bmatrix} \xrightarrow{rank} \begin{bmatrix}1&4&5&-1&-2&|0 \\ 0&-9&-9&3&5&|0 \\ 0&0&0&-1&-3&|0 \\ 0&0&0&0&0&|0\end{bmatrix}$$

$$a_5 = t$$ $$a_4 = -3t$$ $$a_3 = s$$ $$a_2 = 9s+4t$$ $$a_1 = -41s-17t$$

Therefore, the general solution:

$$\Rightarrow t(-17,4,0,-3,1) + s(-41,9,1,0,0)$$

Take t = 1: $$a_5 = 1$$ $$a_4 = -3$$ $$\Rightarrow a_4(x^3-x^2+x+1)+a_5(2x^3-x^2+4x+5)$$ $$\Rightarrow -3(x^3-x^2+x+1)+1(2x^3-x^2+4x+5)$$ $$\Downarrow$$ $$-x^3+2x^2+x+2$$

Is it the only vector in the intersection?

$$dim(U_1 \cup U_2) = dim(U_1) + dim(U_2) - dim(U_1 \cap U_2)$$

I calculated in previous question so: $$dim(U_1 \cap U_2) = 2 + 2 - 3 = 1$$

So there is only 1 vector in the base of $(U_1 \cap U_2)$.

$$B_{(U_1 \cap U_2)} = \{-x^3+2x^2+x+2\}$$

Alon
  • 1,647
  • See the answer to https://math.stackexchange.com/q/2353834/265466. Your question is effectively a duplicate of that one. – amd Dec 27 '19 at 18:58
  • Thank you, I dont know why i didnt find it myself. – Alon Dec 27 '19 at 19:41
  • So if i understand well the link. i can just find a solution for $a_5,a_4$, say for $t=1$ and find the vector that in the intersection... not sure, how do i know that i have only 1 vector in the intersection? – Alon Dec 27 '19 at 19:43
  • Here’s an even better reference. One of the answers gives the general method. Note, though, that whichever of the answers that you follow, you’ll have to adjust for the fact that the vectors that define $U_1$ aren’t linearly independent. – amd Dec 27 '19 at 22:31
  • Thank you, do you know to say if what i did is correct or not? – Alon Dec 27 '19 at 23:25

2 Answers2

1

I’d word a few things differently, but your solution looks correct to me.

Note that you could’ve found bases for both $U_1\cap U_2$ and $U_1+U_2$ (which you asked about in a previous question) at the same time. Forming the combined matrix of all of the spanning vectors and fully reducing that matrix, we have $$\left[\begin{array}{ccc|cc}1&4&5&1&2\\2&-1&1&-1&-1\\3&3&6&1&4\\6&6&12&1&5\end{array}\right] \to \left[\begin{array}{ccc|cc}1&0&1&0&\frac79\\0&1&1&0&-\frac49\\0&0&0&1&3\\0&0&0&0&0\end{array}\right].$$ Examining the pivots of the reduced matrix as a whole, we can see that a basis for $U_1+U_2$ is $\{(1,2,3,6)^T,(4,-1,3,6)^T,(1,-1,1,1)^T\}$. This is no doubt a different basis than the one that you obtained by row-reducing the transpose of this matrix. Note, too, that this reduced matrix tells us that $\dim(U_1)=\dim(U_2)=2$, so just as you’ve calculated, we must have $\dim(U_1\cap U_2)=1$.

For $U_1\cap U_2$, we examine the null space, which is spanned by $(1,1,-1,0,0)^T$ and $(7/9,-4/9,0,3,-1)^T$. The first of these only tells us that the generating vectors of $U_1$ aren’t linearly independent, which doesn’t really help. On the other hand, the second one tells us that $$\frac79\begin{bmatrix}1\\2\\3\\6\end{bmatrix}-\frac49\begin{bmatrix}4\\-1\\3\\6\end{bmatrix} = -3\begin{bmatrix}1\\-1\\1\\1\end{bmatrix}+1\begin{bmatrix}2\\-1\\4\\5\end{bmatrix}.$$ This vector, which corresponds to $-x^3+2x^2+x+2$, lies in the intersection of the two spaces, which we know from above is one-dimensional, so that’s our basis for the intersection.

amd
  • 53,693
0

Note that the sum of the third vector spanning $U_1$ is the sum of the first two. Taking (4 x 1st vector - 2nd vector)/9 we get $u_1=x^2+x+2$. Subtracting $4u_1$ from the 1st vector gives $u_2=x^3-2x^2-x-2$, so $u_1,u_2$ span $U_1$.

For $U_2$ subtracting twice the 1st vector from the second gives $v_1=x^2+2x+3$. Subtracting that from the first vector gives $u_2$, so $U_2$ is spanned by $v_1,u_2$.

Clearly, multiples of $u_2$ belong to $U_1\cap U_2$. So $U_1\cap U_2$ has dimension 1 or 2. But $u_2+u_1=x^3-x^2\in U_1$ and $x^3-x^2\notin U_2$, so $U_1\cap U_2$ must have dimension 1 and hence be just the multiples of $u_2$.

almagest
  • 18,380
  • I got minus the value you got for the intersecting vector, do you see any mistake in my way?

    and i didnt realy understood your way, you guessed?

    – Alon Dec 27 '19 at 20:23
  • Any scalar multiple of a generating vector also generates! – almagest Dec 27 '19 at 20:28
  • Yes that i have already understood, thats why i thought we got to the same answer.

    But, i worked "hard" technically to get to the answer and you like said: multiply this by this and substract this... and here you have it. ??? this looks like guesses, did you just guess?

    – Alon Dec 27 '19 at 20:42
  • There are many ways of answering any given problem. All that matters is that the answer is (1) correct, and (2) clear. If you happen to notice particular features of the problem, then a "non-standard" approach may be quicker. – almagest Dec 27 '19 at 20:46