What extra condition must be satisfied by the kernel to find the unknown function in this system of equations? How?

46 Views Asked by At

There are two integral equations of $$\alpha(x,y)=\int_x^y K(x,y,t)U(t)dt$$ $$\beta(x,y)=\int_x^y K(x,y,t)U(t)tdt$$

where $K(x,y,t)$ is a non-negative known function which $t$ lives in $[y,x]$; $U(t)$ is an unknown function in $(-∞,∞)$ that is differentiable; $U(0)=0$; $α(x,y)$ and $β(x,y)$ are known; $x$ and $y$ are independent. What condition must be satisfied by $K(x,y,t)$ in order to find $U(t)$, or $U(α)$ and $U(β)$? Analytically or numerically?

Otherwise, if only a list of data points $(x,y,α,β)$ is given, can $U(t)$ still be found?

I am not a math student and have little knowledge of math. I tried to read a book while I noticed this may relate to the field of Integral Equations but it is too difficult for me to understand a bit. I would like to know if there is a situation these equations are solvable and what particular method can I use. In fact, the second case is the one in real life.

If for certain case (eg. polynomials), $U(t)$ can be solved, I am thinking maybe I can use something like the Taylors Series to approximate the $K(x,y,t)$.