3

Drawing the graph of a function is not intuitive. For example, the function $f(x)=x^2$ is special, we can say that given two points of the function the straight line that join those points is above the function. So we write $$ x_2<(x_1+x_0)(x-x_0)+x^2 $$ How did I obtain the part of the right? $$ (Y-x_0^2)/ (x-x_0) = (x_1^2 - x_0^2) / (x_1 - x_0) $$ Now I just isolate Y in the equation above.

My question is, can I do the same with $f(x)=x^3$? And how do I prove it? Can I say that $x^3$ is below the straight line that join two points of the function $x^3$? If and only if $x$ is positive.

Victory
  • 39
  • I'm not sure exactly what you are asking, but...if you draw the line from $(-1,(-1)^3)$ to $(1,1^3)$, can you show that there are portions of the curve that lie both above and below that line? – rogerl Sep 04 '18 at 00:08
  • Sorry. My english sometimes only sucks. I can prove that by deriving the function... but what I want is doing the same without it. – Victory Sep 04 '18 at 00:13

2 Answers2

3

Hint:   for $\,x_{0,1} \ge 0\,$ the generalized means inequality gives $\,\sqrt[3]{\dfrac{x_0^3+x_1^3}{2}} \ge \dfrac{x_0+x_1}{2}\,$, which is equivalent to $\,f(x)=x^3\,$ being midpoint-convex on $\,\Bbb R^+\,$. By continuity, it follows that $\,f(x)\,$ is in fact convex on $\,\Bbb R^+\,$. More generally, the same argument works for $\,f(x)=x^n\,$, $\,n \ge 1\,$.

[ EDIT ]   Since $\,f(x)=x^3\,$ is an odd function, it follows that it is concave on $\,\Bbb R^-\,$. Since the direction of concavity changes at $0$, the function is neither convex nor concave on the entire $\,\Bbb R^\,$.
dxiv
  • 76,497
  • Thanks!! Your are right. :) that's a new point of view. – Victory Sep 04 '18 at 00:32
  • But can I do the same only with the definition of the straight rect and de function of x^3? – Victory Sep 04 '18 at 02:39
  • @Josselyn do the same Do you mean determine convexity on the half-axes, or prove it is not convex on the entire axis, or something else? only with Not sure what you mean to exclude from being used in a proof by that only. – dxiv Sep 04 '18 at 03:17
1

For a graph of a function to be convex the epigraph must be a convex set. This means I can take any two points in the set and draw a straight line between them without leaving the set. To see that this isn't true for $f(x)=x^3$ take any $x < 0$ and draw the line from $(x,f(x))$ to the point $(0,0)$ which are both in the epigraph of the function, but none of the points between them are in the epigraph so $f(x)$ cannot be convex. The concave case is the same but with $x>0$.

CyclotomicField
  • 11,018
  • 1
  • 12
  • 29
  • That's right. So I can take two different cases. Whenever I take x<0 or x>0. So, now I have to prove that x3>(x1+x0)(x-x0)+x^2 if x is negative and the other case when x is positive. – Victory Sep 04 '18 at 00:23
  • And I don't know where do I start. Because my domain is x€[x0,x1] – Victory Sep 04 '18 at 00:24
  • 2
    @Josselyn you only need two points as counter-examples and if you consider $x=\pm 1$ then consider the linear segment connecting it to with $(0,0)$. This is simply $y=x$ on both sides of the origin. What can you tell me about $y=(-\frac{1}{2})^3$ versus $y=-\frac{1}{2}$? – CyclotomicField Sep 04 '18 at 02:42