Given a constant vector $\mathbf{c}$ and a vector of variables $\mathbf{x}$, what can we say about the equation:


where $\langle*,*\rangle$ represents the dot product and $s$ is a real number.

This is a sum of $c_ix_i$ terms and since we have only one equations and more unknowns we have an infinity of solutions. However, how can I say anything about this equation?

I tried to use the representation of the dot product $\langle\mathbf{c},\mathbf{x}\rangle=\|\mathbf{c}\|\|\mathbf{x}\|\cos\theta$, where $\theta$ is the angle between the vectors.

$\langle\mathbf{c},\mathbf{x}\rangle=\|\mathbf{c}\|\|\mathbf{x}\|\cos\theta=s$ $\implies \|\mathbf{x}\|=\frac{s}{\|\mathbf{c}\|\cos\theta}$

Is this the most we can say about this equation and its "solution"? For example, we can see that $\|\mathbf{x}\|$ will be in the interval $[\frac{s}{\|\mathbf{c}\|},\frac{s}{\|\mathbf{c}\|}]$.

What else? Can we, for example, find a bound for $\langle\mathbf{c},\mathbf{x}\rangle$ ?

  • $\begingroup$ What are you looking for? $\endgroup$ – Alex Silva Dec 2 '14 at 23:48
  • $\begingroup$ If the equation is $\langle\bf{c, x}\rangle = s$ then I'd say the bound for $\langle\bf{c, x}\rangle$ is $s$. $\endgroup$ – Jim Dec 3 '14 at 0:00

Thinking geometrically, $\langle\mathbf{c},\mathbf{x}\rangle=s$ is equivalent to

$$\langle\frac{\mathbf{c}}{||\mathbf{c}||},\mathbf{x}\rangle= \frac{s}{||\mathbf{c}||}$$

This then is satisfied by all vectors $\mathbf{x}$ that have a projection of $\displaystyle \frac{s}{||\mathbf{c}||}$ in direction $\displaystyle \hat{\mathbf{c}} = \frac{\mathbf{c}}{||\mathbf{c}||}$. Thus you can write $\mathbf{x}$ as

$$\mathbf{x} = \frac{s}{||\mathbf{c}||} \hat{\mathbf{c}} + \mathbf{w}$$

where $\mathbf{w}$ is any vector orthogonal to $\hat{\mathbf{c}}$. That is, the locus of such $\mathbf{x} $ is a plane with normal $\hat{\mathbf{c}}$.


Let $c_1$ be the unit vector along $c$ direction, let $I$ be the unit matrix. Then the vector $w$ in the previous answer can be expressed as $$w=b.(I-c_1c_1)$$. Where $b$ is an arbitrary vector.

Thus $w.c=0$.

Sorry, I typed everything from my iPhone, which is not convenient.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.