$f\in C^1$ convex, if exists $x^*\in S$ such that $\nabla^t f(x^*)(y-x^*)\ge 0$ then $x^*$ is a global minimizer of $f$

28 Views Asked by At

Let $f\in C^1$ convex defined in a convex $S$. If there exists $x^*\in S$ such that for all $y\in S$ it is true that $$\nabla^t f(x^*)(y-x^*)\ge 0$$ then $x^*$ is a global minimizer of $f$ in $S$

I know from here $f$ convex $\iff$ $f(y)\ge f(x)+\nabla f(x)(x-y)$ so we know that for this particular $x^*$ and any $y$ we have

$$f(y)\ge f(x^*) + \nabla f(x^*)(y-x^*) \implies f(y)-f(x^*)\ge \nabla f(x^*)(y-x^*)\ge 0\implies f(y)\ge f(x^*)$$

for any $y$, which means $x^*$ is a global minimizer.

Can you spot any errors in my proof?