I want to know if there is a general rule that will give me the answer. I'm not talking about crazy expressions under a radical, just simple variables raised to a fractional exponent like:
$$X^{4/3}-5X^{2/3}+6=0$$
I want to know if there is a general rule that will give me the answer. I'm not talking about crazy expressions under a radical, just simple variables raised to a fractional exponent like:
$$X^{4/3}-5X^{2/3}+6=0$$
On
There is no "general rule" for solving all such equations. But your equation is a quadratic equation in disguise. Such questions mostly appear in pre-calculus exercises.
A rule of thumb is that you can make your equation easier by introducing new variables.
Since $x^{m/n}=(x^{1/n})^m$, you can write $$ (x^{4/3})=(x^{1/3})^4,\quad (x^{1/3})^2 $$ so that your equation becomes $$ (x^{1/3})^4-5(x^{1/3})^2+6=0\ \tag{1} $$
Now if you introduce a new variable $y=x^{1/3}$, your equation becomes $$ y^4-5y^2+6=0\tag{2} $$ which looks simpler since there is no "radical" (fraction powers) anymore.
But this is still not easy: it is of power $4$.
But if you observe that $y^4=(y^2)^2$, (2) is the same as $$ (y^2)^2-5y^2+6=0\tag{3} $$ where if you introduce another new variable $w=y^2$, you have a quadratic equation $$ w^2-5w+6=0 $$ which can be solved easily since $w^2-5w+6=(w-2)(w-3)$.
On
Say that in your equation, the exponents of your unknown $x$ are fractions $$\frac{m_1}{n_1},\ldots,\frac{m_k}{n_k}.$$
Let $N$ be the least common multiple of $n_1,\ldots,n_k$, and let $N = n_ic_i$. Then $$\frac{m_i}{n_i} = \frac{m_ic_i}{n_ic_i} = \frac{m_ic_i}{N}.$$ That means that you can rewrite your equation as an equation in which every radical is an $N$th root. Letting $y=x^{1/N}$, you can then turn your equation into an equation in $y$ which has no radicals, just integral powers. That equation will be an equation of degree $M=\max\{c_1m_1,\ldots, c_km_k\}$, and so has at most $M$ solutions for $y$. When $N$ is odd, each solution for $y$ will give you a unique solution for $x$, yielding at most $M$ solutions for $x$.
If $N$ is even then at least one of the $n_i$ is even, which means that your $x$ must be nonnegative. So nonnegative values of $y$ will give one corresponding solution, but negative values of $y$ will not correspond to solutions of the original equation. In any case, there is at most one $x$ value per $y$-value, so again you get at most $M$ solutions for $x$.
For a case like this we can define $y=x^{\frac 23}$ and make the equation $y^2-5y+6=0$ which has two roots in $y$. Then you can have either square root, so there will be four zeros. You need to carefully specify the range of equations that is acceptable to get any simple rule.