# Darboux integration, show inequality

Suppose f, g:[a,b]->R are bounded & g(x)<=f(x) for all x in [a,b]
for P a partition of [a,b], show that L(g,P)<=L(f,P)

I don't know whether I should show by cases since I don't know the monotonicity of the both functions f and g. It seems like that the graphs of both functions have to behave the same way from the given condition g(x)<=f(x), isn't it? Or else how would I compare g(Xi) with f(Xi-1) or g(Xi-1) with f(x) though?

Case 1) suppose both f, g are monotone increasing
Case 2) suppose both f, g are monotone decreasing
Case 3) suppose f is monotone increasing and g is monotone decreasing
Case 4) suppose f is monotone decreasing and g is monotone increasing

Proof:
Since f, g are bounded, therefore there exists upper and lower integrals.
L(g,P)=sum g(Xi-1)(change of Xi) or L(g,P)=sum g(Xi)(change of Xi)
L(f,P)=sum f(Xi-1)(change of Xi) or L(f,P)=sum f(Xi)(change of Xi)
Since change of Xi are the same, we only need to compare sum g(Xi-1) with sum f(Xi-1) or sum g(Xi) with sum f(Xi).
Since g(x)<=f(x), it follows that sum g(Xi-1) <= sum f(Xi-1) or sum g(Xi) <= sum f(Xi)
Hence L(g,P) <= L(f,P)

Does it make sense? Any comments?

Related Calculus and Beyond Homework Help News on Phys.org

Can you first say what L(f,P) is supposed to be?

L(f,P) is the Lower Darboux Sum; L(f,P)=sum mi(Xi-Xi-1)
U(f,P) is the Upper Darboux Sum; U(f,P)=sum Mi(Xi-Xi-1)
f: is the function
P: is the partition of the domain [a,b]; P={Xo, X1, ... , Xn}
mi: is the greatest lower bound [Inf(f)]
Mi: is the least upper bound [Sup(f)]
delta Xi=Xi-Xi-1: each partition interval of domain [a,b] where Xo=a and Xn=b

Last edited:

I don't think you need monotonicity of any function: take a partition

$$P=\{x_0,...,x_n\}$$

We know that $$f(x)\leq g(x), \forall x\in[a,b]$$. So since f and g are bounded, it follows that

$$m_{i,f}=\inf_{x\in[x_{i-1},x_i]}{f(x)}\leq \inf_{x\in[x_{i-1},x_i]}{f(x)}=m_{i,g}$$

Therefore, it is obvious that

$$m_{i,f}(x_i-x_{i-1})\leq m_{i,f}(x_i-x_{i-1})$$

Taking sums of both sides, yield

$$L_{f,P}=\sum_{i=0}^n{m_{i,f}(x_i-x_{i-1})}\leq\sum_{i=0}^n{m_{i,f}(x_i-x_{i-1})}=L_{g,P)$$

Yeah, I thought so too, just need second opinion. Thanks!