If f(x 1, x2, x3, ... xn)>=f 1(x 1, x2, x3, ... xn)
f 1(x 1，x2，x3，...xn)>=f2(x 1，x2，x3，...xn)
...
fk(x 1，x2，x3，...xn)>=0
Then * holds, and these inequalities are easier to prove than *.
This is the scaling method, which is very simple by using the transitivity of inequality: a >；; =b，b & gt=c
= & gta & gt=c
So ... when an inequality seems difficult to prove, it can be "decomposed" into several steps to prove it.
Disadvantages: easy to cause: excessive scaling
For example, to get a & gt=c
Then the first certificate: A> = B.
But if b & gt=c is not necessarily true, let alone B.
That is failure. .
Therefore, there are two points in practicing the scaling method:
(1) Scales a face into a familiar structure, such as scaling asymmetry into symmetry, scaling uneven times into uniformity, and scaling those that cannot be summed by splitting terms into those that can be summed by splitting terms. . .
(2) Don't exaggerate (this requires experience)
That's all. It's easier said than done. . . Everyone should look at the problems themselves and understand them well.