Chapter 6 Series
6.1 Finite series
Definition 6.1 (Finite series) Let m,n be integers, and let (an)∞n=m be a finite sequence of real numbers, assigning a real number ai to each integer i between m and n inclusive (i.e., m≤i≤n). Then we define the finite sum (or finite series) ∑ni=mai by the recursive formula n∑i=mai:=0 whenever n<m;n+1∑i=mai:=(n∑i=mai)+an+1 whenever n≥m−1.
Strictly speaking, a series is an expression of the form ∑ni=mai; this series if mathematically (but not semantically) equal to a real number, which is the sum of that series. Note that the variable i (sometimes called the index of summation) is a bound variable (sometimes called a dummy variable); the expression ∑ni=mai does not actually depend on any quantity named i.
Lemma 6.1
- Let m≤n≤p be integers, and let ai be a real number assigned to each integer m≤i≤p. Then we have n∑i=mai+p∑i=n+1ai=p∑i=mai.
- Let m≤n be integers, k be another integer, and let ai be a real number assigned to each integer m≤i≤n. Then we have n∑i=ma=n+k∑i=m+kai−k.
- Let m≤n be integers, and let ai,bi be real numbers assigned to each integer m≤i≤n. Then we have n∑i=m(ai+bi)=n∑i=mai+n∑i=mbi.
- Let m≤n be integers, and let ai be a real number assigned to each integer m≤i≤n, and c be another real number. Then we have n∑i=m(cai)=cn∑i=mai.
- (Triangle inequality fir finite series) Let m≤n be integers, and let ai be a real number assigned to each integer m≤i≤n. Then we have |n∑i=mai|≤n∑i=m|ai|.
- (Comparison test for finite series) Let m≤n be integers, and let ai,bi be real numbers assigned to each integer m≤i≤n. Suppose that ai≤bi for all m≤i≤n. Then we have n∑i=mai≤n∑i=mbi.
Definition 6.2 (Summations over fintite sets) Let X be a finite et with n elements (where n∈N), and let f:X→R be a function from X to the real numbers. Then we can define the finite sum ∑x∈Xf(x) as follows. We first select and bijection g from {i∈N:1≤i≤n} to X; such a bijection exists since X is assumed to have n elements. We then define ∑x∈Xf(x):=n∑i=1f(g(i)).
Proposition 6.1 (Finite summations are well-defined) Let X be a finite et with n elements (where n∈N), and let f:X→R be a function, and let g:{i∈N:1≤i≤n}→X and h:{i∈N:1≤i≤n}→X be bijections. Then we have n∑i=1f(g(i))=n∑i=1f(h(i)).
Proposition 6.2 (Basic properties of summation over finite sets)
- If X is empty, and f:X→R is a function, we have ∑x∈Xf(x)=0.
- If X consists of a single element, X={x0}, and f:X→R is a function, we have ∑x∈Xf(x)=f(x0).
- (Substitution, part I) If X is a finite set, f:X→R is a function, and g:Y→X is a bijection, then ∑x∈Xf(x)=∑y∈Yf(g(y)).
- (Substitution, part II) Let n≤m be integers, and let X be the set X:={i∈Z:n≤i≤m}. If ai is a real number assigned to each integer i∈X, then we have m∑i=nai=∑i∈Xai.
- Let X,Y be disjoint finite sets, and f:X∪Y→R is a function. Then we have ∑z∈X∪Yf(z)=∑x∈Xf(x)+∑y∈Yf(y).
- (Linearity, part I) Let X be a finite set, and let f:X→R and g:X→R be functions. Then ∑x∈X(f(x)+g(x))=∑x∈Xf(x)+∑x∈Xg(x).
- (Linearity, part II) Let X be a finite set, and let f:X→R be a function. Then ∑x∈Xcf(x)=c∑x∈Xf(x).
- (Monotonicity) Let X be a finite set, and let f:X→R and g:X→R be functions such that f(x)≤g(x) for all x∈X. Then we have ∑x∈Xf(x)≤∑x∈Xg(x).
- (Triangle inequality) Let x be a finite set, and let f:X→R be a function, then |∑x∈Xf(x)|≤∑x∈X|f(x)|.
The substitution rule in Proposition 6.2 indicates that we can rearrange the elements of a finite sequence at will and still obtain the same value.
Lemma 6.2 Let X,Y be a finite set, and let f:X×Y→R be a function. Then ∑x∈X(∑y∈Yf(x,y))=∑(x,y)∈X×Yf(x,y).
Corollary 6.1 (Fubini's theorem for finite series) Let X,Y be finite sets, and let f:X×Y→R be a function. Then ∑x∈X(∑y∈Yf(x,y))=∑(x,y)∈X×Yf(x,y)=∑(y,x)∈Y×Xf(x,y)=∑y∈Y(∑x∈Xf(x,y))
Exercise 6.1 Let X be a finite set, let m be an integer and for each x∈X let (an(x))∞n=m be a convergent sequence of real numbers. Show that the sequence ∑x∈X(an(x))∞n=m is convergent, and lim
6.2 Infinite series
Definition 6.3 (Formal infinite series) A (formal) series is any expression of the form \sum_{n=m}^{\infty}a_n,
where m is an integer, and a_n is a real number for any integer n\geq m.
Definition 6.4 (Convergence of series) Let \sum_{n=m}^{\infty}a_n be a formal infinite series. For any integer N \geq m, we define the N^{\text{th}} partial sum S_N of this series to be S_N:=\sum_{n=m}^Na_n. If the sequence (S_N)^{\infty}_{N=m} converges to some limit L as N\to\infty, then we say that the infinite series \sum_{n=m}^{\infty}a_n is convergent, and converges to L; we also write L=\sum_{n=m}^{\infty}a_n, and say that L is the sum of the infinite series \sum_{n=m}^{\infty}a_n. If the partial sums S_N diverge, then we say that the infinite series \sum_{n=m}^{\infty}a_n is divergent, and we do not assign any real number value to that series.
Proposition 6.3 Let \sum_{n=m}^{\infty}a_n be a formal series of real numbers. Then \sum_{n=m}^{\infty}a_n converges iff for every real number \varepsilon>0. there exists an integer N\geq m such that \left| \sum_{n=p}^{q}a_n\right|\leq \varepsilon \text{ for all }p,q\geq N
Corollary 6.2 (Zero test) Let \sum_{n=m}^{\infty}a_n be a convergent series of real numbers. Then must have \lim_{n\to\infty}a_n. To put this another way, if \lim_{n\to\infty}a_n is non-zero or divergent, then the series \sum_{n=m}^{\infty}a_n is divergent.
Definition 6.5 (Absolute convergence) Let \sum_{n=m}^{\infty}a_n be a formal series of real numbers. We say that this series is absolutely convergent iff the series \sum_{n=m}^{\infty}|a_n| is convergent.
In order to distinguish convergence from absolute convergence, we sometimes refer to the former as conditional convergence.
Proposition 6.4 (Absolute convergence test) Let \sum_{n=m}^{\infty}a_n be a formal series of real numbers. If this series is absolutely convergent, then it is also conditionally convergent. Furthermore, in this case we have the triangle inequality \left|\sum_{i=m}^{\infty}a_i\right|\leq\sum_{i=m}^{\infty}|a_i|.
Proposition 6.5 (Alternating series test) Let \sum_{n=m}^{\infty}a_n be a sequence of real numbers which are non-negative and decreasing. Then the series \sum_{n=m}^{\infty}(-1)^na_n is convergent iff the sequence a_n converges to 0 as n\to\infty.
Proposition 6.6 (Series laws)
- If \sum_{n=m}^{\infty}a_n is a series of real numbers converging to x, and \sum_{n=m}^{\infty}b_n is a series of real numbers converging to y, then \sum_{n=m}^{\infty}(a_n+b_n) is also a convergent series, and converges to x+y. In particular, we have \sum_{i=m}^{\infty}(a_i+b_i)=\sum_{i=m}^{\infty}a_i+\sum_{i=m}^{\infty}b_i.
- If \sum_{n=m}^{\infty}a_n is a series of real numbers converging to x, and c is a real number, then \sum_{n=m}^{\infty}(ca_n) is also a convergent series, and converges to cx. In particular, we have \sum_{i=m}^{\infty}(ca_i)=c\sum_{i=m}^{\infty}a_i.
- Let \sum_{n=m}^{\infty}a_n be a series of real numbers, and k\geq 0 bean integer. If one of the two series \sum_{n=m}^{\infty}a_n and \sum_{n=m+k}^{\infty}a_n are convergent, then the other one is also, and we have the identity \sum_{n=m}^{\infty}a_n=\sum_{n=m}^{m+k-1}a_n+\sum_{n=m+k}^{\infty}a_n
- Let \sum_{n=m}^{\infty}a_n be a series of real numbers converging to x, and let k be an integer. Then \sum_{n=m+k}^{\infty}a_{n-k} also converges to x.
Lemma 6.3 (Telescoping series) Let (a_n)_{n=0}^{\infty} be a sequence if real numbers which converge to 0. Then the series \sum_{n=0}^{\infty}(a_n-a_{n+1}) converges to a_0.
6.3 Sums of non-negative numbers
Proposition 6.7 Let \sum_{n=m}^{\infty}a_n be a formal series of non-negative real numbers. Then this series is convergent iff there is a real number M such that
\sum_{n=m}^{N}a_n\leq M \text{ for all integers } N\geq m.
:::{corollary name=“Comparision test”} Let \sum_{n=m}^{\infty}a_n and \sum_{n=m}^{\infty}b_n be two formal series of real numbers, and suppose that |a_n|\leq b_n for all n\geq m. Then if \sum_{n=m}^{\infty}b_n is convergent, then \sum_{n=m}^{\infty}a_n is absolutely convergent, and in fact
\left|\sum_{n=m}^{\infty}a_n\right|\leq \sum_{n=m}^{\infty}|a_n|\leq\sum_{n=m}^{\infty}b_n. :::
:::{.lemma name=“Geometric series} Let x be a real number. If |x|\geq 1, the the series \sum_{n=0}^{\infty}x^n is divergent. If however |x|<1, then the series is absolutely convergent and \sum_{n=0}^{\infty}x^n = 1/(1-x). :::
Let (a_n)_{n=1}^{\infty} be a decreasing sequence of non-negative real numbers (so a_n\geq 0 and a_{n+1}\leq a_n for ll n\geq 1). Then the series \sum_{n=1}^{\infty}a_n is convergent iff the series \sum_{k=0}^{\infty}2^ka_{2^k}=a_1+2a_2+4a_4+\dots is convergent.
Let q>0 be a rational number. Then the series \sum_{n=1}^{\infty}1/n^q is convergent when q>1 and divergent when q\leq 1.
Then quantity \sum_{n=1}^{\infty}1/n^q, when it converges, is called \zeta(q), then Riemann-zeta function of q.
6.4 Rearrangement of series
Proposition 6.8 Let \sum_{n=0}^{\infty}a_n be a convergent series of non-negative real numbers, and let f:\mathbb{N}\to\mathbb{N} be a bijection. Then \sum_{m=0}^{\infty}a_{f(m)} is also convergent, and has the same sum: \sum_{n=0}^{\infty}a_n=\sum_{m=0}^{\infty}a_{f(m)}.
Proposition 6.9 (Rearrangement of series) Let \sum_{n=0}^{\infty}a_n be an absolutely convergent series o real numbers, and let f:\mathbb{N}\to\mathbb{N} be a bijection. Then \sum_{m=0}^{\infty}a_{f(m)} is also absolutely convergent, and has the same sum: \sum_{n=0}^{\infty}a_n=\sum_{m=0}^{\infty}a_{f(m)}.
A series which is conditionally convergent but bot absolutely convergent can in fact be rearranged to converge to any value.
6.5 The root and ratio tests
Theorem 6.1 (Root test) Let \sum_{n=m}^{\infty}a_n be a series of real numbers, and let \alpha:=\limsup_{n\to \infty}|a_n|^{1/n}.
- If \alpha<1, then the series \sum_{n=0}^{\infty}a_n is absolutely convergent (and hence conditionally convergent).
- If \alpha>1, then the series \sum_{n=0}^{\infty}a_n is not conditionally convergent (and hence absolutely convergent).
- If \alpha=1, we cannot assert any conclusion.
Lemma 6.4 Let (c_n)_{n=m}^{\infty} be a sequence of positive numbers. Then we have \liminf_{n\to\infty}\frac{c_{n+1}}{c_n} \leq \liminf_{n\to\infty}c_n^{1/n} \leq \limsup_{n\to\infty}c_n^{1/n} \leq \limsup_{n\to\infty}\frac{c_{n+1}}{c_n}.
Corollary 6.3 (Ratio test) Let \sum_{n=m}^{\infty}a_n be a series of non-zero real numbers, and let \alpha:=\limsup_{n\to \infty}\frac{|a_{n+1}|}{|a_n|}.
- If \alpha<1, then the series \sum_{n=0}^{\infty}a_n is absolutely convergent (and hence conditionally convergent).
- If \alpha>1, then the series \sum_{n=0}^{\infty}a_n is not conditionally convergent (and hence absolutely convergent).
- If \alpha=1, we cannot assert any conclusion.
We have \lim_{n\to \infty}n^{1/n}=1.