Prove convexity of complicated rational function

755 Views Asked by At

Can anyone help me prove the convexity of this rational function? The man who proved the convexity of function used these facts. But I don't know this fact is correct or not. Here are the facts and function:

  1. As N increases, f(N) goes to infinity, That implies that there must be a minima (either at N=0 or somewhere else with a finite N)
  2. There cannot be more than one (positive) minima since we're dealing with second order equation.

f(N) = +-cN^4+-dN^3+-eN^2+-fN+-g / +-aN^2+-bN

a,b,c,d,e,f,g is constants. and N >=1. I guess the second order equation means that between the leading coefficient of the numerator and the denominator is 2.

Are these facts correct? I think fact 1 is no problem, but fact 2 is correct or not.

I am waiting for any answers. thank you.

1

There are 1 best solutions below

2
On BEST ANSWER

Fact $2$ seems incorrect (but I may not be understanding your notation). Local minima correspond to points where the derivative vanishes. For a rational function of the form $p/q$ where $p$ is of degree $n$ and $q$ is of degree $m$, the numerator of the derivative is a polynomial of degree $n+m-1$ and as such may have up to $n+m-1$ roots, which lead to at most $n+m-1$ possible local minima. In this case, there may be as many as $5$ local minima.