Midpoint convexity及其推广

2019-05-05  本文已影响0人  坐看云起时zym
Midpoint convexity.png

先补充一下闭集(closed set)的定义:集合C为闭集当且仅当C包含C的所有极限点(limit points)。具体可见https://en.wikipedia.org/wiki/Closed_set

证明如下:
即证:\forall a,b \in C, \theta \in [0,1], \theta a+ (1 - \theta)b \in C
x = \theta a+ (1 - \theta)b,a_{0} = a, b_{0} = b
依据如下定义,构造\left \{ a_{n} \right \}, \left \{ b_{n} \right \}:
(i)若x在线段l_{[\frac{a_{n-1} + b_{n-1}}{2},b_{n-1}]}上,则a_{n} = \frac{a_{n-1} + b_{n-1}}{2}, b_{n} = b_{n-1}
(ii)若x在线段l_{[a_{n-1},\frac{a_{n-1} + b_{n-1}}{2}]}上,则a_{n} = a_{n-1}, b_{n} = \frac{a_{n-1} + b_{n-1}}{2}
显然,\left \| b_{n} - a_{n} \right \|_{2} = \frac{\left \| b - a \right \|_{2}}{2^{n}}
注意到,\forall n \in N, a_{n} \leqslant x \leqslant b_{n},即\left \|x - a_{n} \right \|_{2} \leqslant \left \| b_{n} - a_{n} \right \|_{2}
\lim_{n\rightarrow \infty } \left \|x - a_{n} \right \|_{2} \leqslant \lim_{n\rightarrow \infty } \left \| b_{n} - a_{n} \right \|_{2} = 0
\therefore x = \lim_{n\rightarrow \infty }a_{n} = \lambda
\because a_{n} \in C, C为闭集(closed set), \therefore \lambda \in C, \therefore x \in C #得证

上一篇下一篇

猜你喜欢

热点阅读