0

Can every integration and differentiation on $\mathbb{R}^2$ be determined exactly?

I am curious of this, because I know that there are some integration and differentiation that do not yet have a way to solve them.

However, I also never heard of any theorem that state that there are some integrals and differentiations that cannot be solved.

So, is there any theorem?

Kaveh
  • 3,572
Tao Mao
  • 35

1 Answers1

0

I don't know where to find the proof, but if you restrict your domain to $\mathbb{R}$, then the following equation is known to have no solution in terms of elementary functions. $$\int e^{x^2}dx$$ In addition, the elliptic integrals (arc-length of an ellipse) do not necessary have solutions in terms of elementary functions. For instance: $$L=\int \sqrt{1+\left(\frac{dy}{dx}\right)^2}dx$$ Take the equation of an ellipse: $$ax^2+by^2=c^2$$ Where $a\neq b$. If $a=b$, then you have a circle, which has a trivial arc length. Use implicit differentiation to solve: $$2axdx+2bydy=0 \rightarrow axdx=-bydy \rightarrow \frac{dy}{dx}=\frac{-ax}{by}$$ Use the first quadrant, so: $$y=\sqrt{\frac{c^2-ax^2}{b}}$$ So combining equations gives: $$L=\int \sqrt{1+\left(\frac{-ax}{b\sqrt{\frac{c^2-ax^2}{b}}}\right)^2}$$ $$L=\int \sqrt{1+\frac{a^2x^2}{b(c^2-ax^2)}}dx$$ This is the formula for the arc-length of an ellipse, and it has no elementary solution. It is defined as an elliptic integral of the second kind.