Amount of solutions added to a system of equations through the application of non-invertible operations.

52 Views Asked by At

Let's say we had a linear equation of the form $ax+b=c$ we then solve it for $x$ getting, let's say, $x=5$.

Just for fun, let's pretend we haven't realized we had solved the problem, so we square both sides of this equality getting $x^2=25$, we then try to solve it by square rooting and we get $|x|=|5| \implies x= \pm5$.

If we try this in the original equation, only one of the $x$'s will be a solution.

Generalizing this process, is there any way to know exactly how many "solutions" are we "adding" to any system of equations by applying irreversible operations?

1

There are 1 best solutions below

2
On BEST ANSWER

The general philosophy concerning this question is as follows:

An equation, or system of equations, $$\Phi(x)=0\tag{1}$$ making sense for elements $x$ of an agreed ground set $X$ defines a solution set $$S=\{x\in X\>|\>\Phi(x)=0\}$$ in an implicit way: For each $x\in X$ it is easy to test whether it belongs to $S$ or not, but we do not have an a priori overview over $S$. The solution set $S$ may be empty, it can contain exactly one element, a certain finite number of elements, or infinitely many elements.

Solving the equation or system $(1)$ means producing an explicit description of $S$ in the form of a list, say $S=\{-1, 3,17\}$, or a parametric representation ("production scheme"): $S=\{ x(\iota)\>|\>\iota\in I\}$, whereby $I$ is an index set and $\iota\mapsto x(\iota)\in X$ is an explicitly given function.

In order to solve $(1)$ it is very common to set up a chain of true statements $$\bigl(x\in S\ \Leftrightarrow\bigr)\quad\Phi(x)=0\quad\Rightarrow\ldots\Rightarrow\ldots\Rightarrow\ldots\Rightarrow \quad x\in \hat S\ ,\tag{2}$$ whereby subsequent statements are related by some algebraic manipulation or simplification, and the resulting set $\hat S\subset X$ is an explicitly described set. There is, however, the following caveat: If the individul steps are not all reversible then we have not proven $S=\hat S$, but only $S\subset \hat S$, and we have to check each individual element $x\in\hat S$ whether it actually satisfies the original equation $(1)$.

This check may be omitted only if all steps in the chain $(2)$ are reversible, or if the given equation $(1)$ is an instance of equations for which we have a general theory. If this theory guarantees, e.g., "exactly one solution", and the obtained $\hat S$ is a singleton $\{a\}$ then this $a$ is "the" solution of $(1)$. Similarly, if the general theory guarantees a two dimensional vector space of solutions, and we have found by whatever means two linearly independent solutions $y_1$, $y_2$, then $S=\{y=c_1 y_1+c_2 y_2\>|\>c_1, c_2 \in{\mathbb R}\}$ without further ado.