I want to solve a problem for an increasing function $g(x)$, for $x \in [0,1]$ and with $g(0) = 0$ and $g(1) = 1$.
The solution will be solution to the following equation $\forall x$, $f_1(x) = f_2(g(x))$.
(with $f_1$ and $f_2$ known)

However, $f_2$ may not be invertible (it is piecewise monotone though). Thus for some $x$, you have multiple solution $g(x)$, that we'll denote $(g^1(x), g^2(x), ..., g^k(x))$ (given the structure of the problem, you always have a countable number of solutions).

However, overall, there is a unique increasing solution $g^*()$ of $[0,1] \rightarrow [0,1]$.
My problem is, I don't know how to impose the monotonicity constraint when solving a system, such that I should be able to pin down the unique solution 'directly' (I don't know if that's even possible). Or in other words, I don't know tricks to write the system with monotonicity embedded, solving for the complete function $g()$, instead of solving pointwise $x$ by $x$.

I was trying to build a contraction mapping operator (maybe it's not the way to go), but I need to embed the monotonicity constraint into the contraction operator and I don't know how to do that. (Because taken at a given $x$, our system can yield several solution and thus no contraction mapping is feasible).

PS: Just to be clear, in this particular case I know how to pin down $g()$ entirely exploiting piecewise monotonicity (and thus building piecewise invertibility). My question is really about tricks to include the monotonicity into a system (as an additional equation? constraint?), and if possible use contraction mapping with it.


Your Answer

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

Browse other questions tagged or ask your own question.