Derivatives and its relationship to convex and concave functions
Summary
Suppose that f is twice differentiable on an interval I (open or closed, bounded or unbounded). Then
- f″ for all x∈I ⟺ f is convex in I
- f''(x)≤0 for all x∈I ⟺f is concave in I
- If f''(x)>0 for all x∈I ⇒ f is strictly convex in I (opposite is not necessarily true)
- If f''(x)<0 for all x∈I ⇒ f is strictly concave in I (opposite is not necessarily true)
- Examples:
- f(x)=x^2 , f''(x)=2 and the function is strictly convex on any interval.
- f(x)=-x^2 , f''(x)=-2 and the function is strictly concave on any interval.
- f(x)=x , f''(x)=0 and the function is concave and convex on any interval.
- f(x)=x^4 is strictly convex on any interval but f''(x)>0 is false. f''(x)=12x^2 and f^{''}\left( 0 \right)=0 .
- Technical note: If I is a closed interval \left[ a,b \right] then f need not be differentiable at the boundary point for these results to hold
- Example: If f\left( x \right)=\sqrt{x} with domain \left[ 0,∞ \right) then f'\left( x \right)= \frac{1}{2\sqrt{x}} and f is not differentiable at x=0 . However, f^{''}\left( x \right)=-x^{-3/2}/4<0 for x>0 and f is still strictly concave on \left[ 0,∞ \right) .