Simple Functions

1.16 Definition A complex function ss on a measurable space XX whose range consists of only finitely many points will be called a simple function. Among these are the nonnegative simple functions, whose range is a finite subset of [0,)[0, \infty) . Note that we explicitly exclude \infty from the values of a simple function.

If α1,,αn\alpha_1, \dots, \alpha_n are the distinct values of a simple function ss , and if we set Ai={x:s(x)=αi}A_i = \{x: s(x) = \alpha_i\} , then clearly

s=i=1nαiχAi,s = \sum_{i=1}^{n} \alpha_i \chi_{A_i},

where χAi\chi_{A_i} is the characteristic function of AiA_i , as defined in Sec. 1.9(d).

It is also clear that ss is measurable if and only if each of the sets AiA_i is measurable.

1.17 Theorem Let f:X[0,]f: X \to [0, \infty] be measurable. There exist simple measurable functions sns_n on XX such that

(a) 0s1s2f0 \le s_1 \le s_2 \le \dots \le f .

(b) sn(x)f(x)s_n(x) \to f(x) as nn \to \infty , for every xXx \in X .

PROOF Put δn=2n\delta_n = 2^{-n} . To each positive integer nn and each real number tt corresponds a unique integer k=kn(t)k = k_n(t) that satisfies kδnt<(k+1)δnk\delta_n \le t < (k+1)\delta_n . Define

φn(t)={kn(t)δnif 0t<nnif nt.(1)\varphi_n(t) = \begin{cases} k_n(t)\delta_n & \text{if } 0 \le t < n \\ n & \text{if } n \le t \le \infty. \end{cases} \quad (1)

Each φn\varphi_n is then a Borel function on [0,][0, \infty] ,

tδn<φn(t)tif 0tn,(2)t - \delta_n < \varphi_n(t) \le t \quad \text{if } 0 \le t \le n, \quad (2)

0φ1φ2t0 \le \varphi_1 \le \varphi_2 \le \cdots \le t , and φn(t)t\varphi_n(t) \to t as nn \to \infty , for every t[0,]t \in [0, \infty] . It follows that the functions

sn=φnf(3)s_n = \varphi_n \circ f \quad (3)

satisfy (a) and (b); they are measurable, by Theorem 1.12(d).

///

简单说明
φn(t)\varphi_n(t) 是一个阶梯函数。