Finding Orthogonal Vectors in R4 with Norm 1

AI Thread Summary
The discussion focuses on finding two orthogonal vectors in R4 with a norm of 1, specifically orthogonal to the vectors u, v, and w. Participants suggest using the dot product to create three equations based on the orthogonality conditions, which can be solved for the vector components. It is noted that while there are three equations, they can be expressed in terms of a fourth variable, allowing for multiple solutions. The importance of first finding vectors that meet the orthogonality criteria before normalizing them to a unit length is emphasized. The conversation highlights the mathematical approach to solving the problem while addressing the challenges of working in four dimensions.
sevag00
Messages
80
Reaction score
1

Homework Statement



Find two vectors in R4 of norm 1 that are orthogonal to the vectors u = (2, 1, −4, 0),
v = (−1, −1, 2, 2) and w = (3, 2, 5, 4).

Homework Equations





The Attempt at a Solution



What i did was, i let a vector x = (x1, x2, x3, x4) that has a norm of 1 and orthogonal to u,v and w
i.e. x.u = 0; x.v = 0 and x.w = 0.
This where I'm stuck. Some help would be appreciated. Thanks.
 
Physics news on Phys.org
hi sevag00! :smile:

(try using the X2 button just above the Reply box :wink:)

hint: how would you find a vector (x1,x2,x3) orthogonal to (a,b,c) and (d,e,f) ? :wink:
 
Using dot product. x1.a + x2.b + x3.c = 0. Same for (d, e, f).
 
Yeah, but the chapter only explains the dot product. So, i think the problem should be solved using dot product.
 
Last edited:
ok, in that case you have three simultaneous equations (from the dot products) …

solve them the usual way :smile:
 
3 equations with 4 unknowns. Will this yield a solution?
 
3 equations with 3 unknown ratios :wink:

so yes, you can find the ratios :smile:
 
Ah yes. 4 equations with 4 unknowns.
 
  • #10
tiny-tim said:
how about the cross product ? :wink:

sevag00 said:
Yeah, but the chapter only explains the dot product. So, i think the problem should be solved using dot product.
Further, the cross product is defined only in three dimensions while this problem is in four dimensions.
 
  • #11
Wait a minute. You got me confused. There are 3 equations with 4 unknown. The equation of the norm cannot be included with the other three.
 
  • #12
HallsofIvy said:
Further, the cross product is defined only in three dimensions …

i was talking about three dimensions
sevag00 said:
There are 3 equations with 4 unknown. The equation of the norm cannot be included with the other three.

there are only three unknown ratios

so you can find the ratios
 
  • #13
What ratios? I want to find the components of x.
 
  • #14
x2/x1, x3/x1, x4/x1
 
  • #15
Not getting your approach.
 
  • #16
divide your three dot-product=0 equations by x1

that gives you three equations in x2/x1, x3/x1, and x4/x1

then solve them :smile:
 
  • #17
I think i did it right. I replaced x2/x1 with X x3/x1 with Y and x4/x1 with Z. Then solved 3 equations with 3 unknowns.
 
  • #18
sevag00 said:

Homework Statement



Find two vectors in R4 of norm 1 that are orthogonal to the vectors u = (2, 1, −4, 0),
v = (−1, −1, 2, 2) and w = (3, 2, 5, 4).

Homework Equations





The Attempt at a Solution



What i did was, i let a vector x = (x1, x2, x3, x4) that has a norm of 1 and orthogonal to u,v and w
i.e. x.u = 0; x.v = 0 and x.w = 0.
This where I'm stuck. Some help would be appreciated. Thanks.

You can write the three equations in four unknowns as three equations in three unknowns but with the fourth unknown on the right-hand-side. In other words, you can express three of the x_i as functions of the fourth. For example, you can write the second equation ##-1 x_1 - 1 x_2 + 2 x_3 + 2 x_4 = 0## as ##-1 x_1 - 1 x_2 + 2 x_3 = -2 x_4##, or ##x_1 + x_2 - 2x_3 = 2x_4##. Do the same thing for the third equation; the first already has the right form. Now you can solve for ##x_1, x_2, x_3##; this will give them as formulas involving ##x_4##.

BTW: the method above is a standard way of dealing with such problems; it is used all the time in such fields as linear optimization (linear programming).
 
  • #19
I know that method. But then those equations will become parameters. They will depend on the value of x4 which has infinite solutions.
 
  • #20
sevag00 said:
I know that method. But then those equations will become parameters. They will depend on the value of x4 which has infinite solutions.

Of course; that is the whole point. So, if you want to find two different solutions, just give two different values for ##x_4##---any two values you like, except 0. (Putting ##x_4 = 0## gives the zero vector, which cannot be normalized to length = 1!)
 
  • #21
I want direct solution for the x-es. I'm not going to plug in millions of values to see if the norm is 1.
 
  • #22
put x4 = 1, then normalise it
 
  • #23
sevag00 said:
I want direct solution for the x-es. I'm not going to plug in millions of values to see if the norm is 1.

You have completely misunderstood what some of us have been saying. First find two vectors that satisfy the perpendicularity conditions. THEN normalize them.

For example, if I start with a vector like ##v = (1,2,3,4)## I can get a normalized version as
u = \frac{v}{\text{norm of }v}
If v was perpendicular to some vector w then so is u, and u has norm 1. That's all there is to it.
 

Similar threads

Back
Top