Proof of Jensen‘s inequality

the usual form of Jensen’s inequality

For a given convex function $f$, the below inequality is satisfied:
$$f(t{x}_1+(1-t)x_2)\le tf(x_1)+(1-t)f(x_2)$$

where $t$ satisfies $t\in [0,1]$.
(Note: the inequality is reversed if $f$ is concave，$f(t{x}_1+(1-t)x_2)\ge tf(x_1)+(1-t)f(x_2)$)

Proof1

We might as well define $F(x)$ as:
$$F(x)=f(t{x}_1+(1-t)x)-tf(x_1)-(1-t)f(x)$$

where $x_1$ is a real number and $x\ge x_1$.

Then take the derivative of $F(x)$ with respect to $x$ :
$$\begin{array}{ccc}{F}^{\prime }(x)& =& (1-t)f^{\prime }(t{x}_1+(1-t)x)-(1-t)f^{\prime }(x)\\ & & \\ & =& (1-t)(f^{\prime }(x+t(x_1-x))-f^{\prime }(x))\end{array}$$

where $x+t(x_1-x)\le x$.
According to the properties of convex functions, it can be seen that $f^{\prime }(x+t(x_1-x))\le f^{\prime }(x)$ is satisfied.
So $F^{\prime }(x)\le 0$，then:
$$\begin{array}{ccc}F(x)\le F(x_1)& =& f(t{x}_1+(1-t)x_1)-tf(x_1)-(1-t)f(x_1)\\ & & \\ & =& f(x_1)-f(x_1)\\ & & \\ & =& 0\end{array}$$

$$\begin{array}{ccc}F(x)& \le & 0\\ & & \\ f(t{x}_1+(1-t)x)-tf(x_1)-(1-t)f(x)& \le & 0\\ & & \\ f(t{x}_1+(1-t)x)& \le & tf(x_1)-(1-t)f(x)\end{array}$$

the generalized form of Jensen’s inequality

$$f(t_1{x}_1+t_2{x}_2+...+t_n{x}_n)\le t_{1}f(x_1)+t_{2}f(x_2)+...+t_{n}f(x_n)$$

where $t_1,t_2...t_n$ are nonnegative real number such that ${\sum }_i^n{t}_i=1$.

When $n=2$, the generalized form become to the usual form.

Proof2

Suppose the equation holds when $n=k$, that is, $f(t_1{x}_1+t_2{x}_2+...+t_k{x}_k)\le t_{1}f(x_1)+t_{2}f(x_2)+...+t_{k}f(x_k)$.

Then when $n=k+1$,
t 1 f ( x 1 ) + t 2 f ( x 2 ) + . . . + t k + 1 f ( x k + 1 ) = ( 1 − t k + 1 ) { 1 ( 1 − t k + 1 ) [ t 1 f ( x 1 ) + t 2 f ( x 2 ) + . . . + t k f ( x k ) ] } + t k + 1 f ( x k + 1 ) = ( 1 − t k + 1 ) [ t 1 ( 1 − t k + 1 ) f ( x 1 ) + t 2 ( 1 − t k + 1 ) f ( x 2 ) + . . . + t k ( 1 − t k + 1 ) f ( x k ) ] + t k + 1 f ( x k + 1 ) ≥ ( 1 − t k + 1 ) f ( t 1 ( 1 − t k + 1 ) x 1 + t 2 ( 1 − t k + 1 ) x 2 + . . . + t k ( 1 − t k + 1 ) x k ) + t k + 1 f ( x k + 1 ) \begin{matrix} t_1f(x_1)+t_2f(x_2)+...+t_{k+1}f(x_{k+1}) &=&(1-t_{k+1})\{\frac{1}{(1-t_{k+1})}[t_1f(x_1)+t_2f(x_2)+...+t_kf(x_k)]\}+t_{k+1}f(x_{k+1}) \\ \\ &=&(1-t_{k+1})[\frac{t_1}{(1-t_{k+1})}f(x_1)+\frac{t_2}{(1-t_{k+1})}f(x_2)+...+\frac{t_k}{(1-t_{k+1})}f(x_k)]+t_{k+1}f(x_{k+1}) \\ \\ &\ge &(1-t_{k+1})f(\frac{t_1}{(1-t_{k+1})}x_1+\frac{t_2}{(1-t_{k+1})}x_2+...+\frac{t_k}{(1-t_{k+1})}x_k)+t_{k+1}f(x_{k+1}) \\ \\ \end{matrix} t1​f(x1​)+t2​f(x2​)+...+tk+1​f(xk+1​)​==≥​(1−tk+1​){(1−tk+1​)1​[t1​f(x1​)+t2​f(x2​)+...+tk​f(xk​)]}+tk+1​f(xk+1​)(1−tk+1​)[(1−tk+1​)t1​​f(x1​)+(1−tk+1​)t2​​f(x2​)+...+(1−tk+1​)tk​​f(xk​)]+tk+1​f(xk+1​)(1−tk+1​)f((1−tk+1​)t1​​x1​+(1−tk+1​)t2​​x2​+...+(1−tk+1​)tk​​xk​)+tk+1​f(xk+1​)​

We might as well define $u_1=\frac{t_1}{(1-t_{k+1})}{x}_1+\frac{t_2}{(1-t_{k+1})}{x}_2+...+\frac{t_k}{(1-t_{k+1})}{x}_k$ and $u_2=x_{k+1}$.

From the Proof1, we can get $f((1-t_{k+1})u_1+t_{x+1}{u}_2)\le (1-t_{k+1})f(u_1)+t_{k+1}f(u_2)$.

$$t_{1}f(x_1)+t_{2}f(x_2)+...+t_{k+1}f(x_{k+1})\ge (1-t_{k+1})f(u_1)+t_{k+1}f(u_2)\ge f((1-t_{k+1})u_1+t_{x+1}{u}_2)=f(t_1{x}_1+t_2{x}_2+...+t_n{x}_n)$$

We deduce the inequality is true for $n+1$, and we have proved the correctness for $n=2$, by the principal of mathematical induction it follows that the result is also true for all integer $n$ greater than 2.