Let f(t) be a convex function and define g(t) to be the running average of f(t)
g
(
t
)
=
1
t
∫
0
t
f
(
τ
)
d
τ
g(t) = \frac{1}{t} \displaystyle\int_0^t f(\tau) ~d\tau
g(t)=t1∫0tf(τ) dτ
Then g is convex.
This is easy (not exactly) enough to prove just by differentiating twice.
但是如果 f ( x ) f(x) f(x)不可导的,好像用不了这个方法。
在 math.stackexchange 上看到有人想用convexity preserving operation证明:
Alternate way to prove running average of a convex function is convex?
以及关于这个问题,想要一个几何方面的解释的:
Running average of a convex function is convex
(一)两次求导: