Week 1βContinuity of the inverse function. Infinite series
Version 2025/02/02 Week 1 in PDF All notes in PDF To other weeks
These notes are being developed to reflect the content of the Real Analysis course as taught in the 2024/25 academic year. The first half of the course is lectured by Dr Yuri Bazlov. Questions and comments on these lecture notes should be directed to Yuri.Bazlov@manchester.ac.uk. The second half will be lectured by Dr Mark Coleman.
Pre-requisite: Mathematical Foundations and Analysis
We build upon what was achieved in the Mathematical Foundations and Analysis (MFA) course, taught in Semester 1. Limit of a sequence, continuous function and limit of a function remain
key notions in Real Analysis, which develops the βanalysisβ part of MFA further. Important functions of real variable, used in MFA, will be formally defined, and their properties proved. This includes the power function
Hence, the Real Analysis course will constantly require the students to call on their knowledge of definitions and results introduced in MFA.
An informal introduction (not covered in lectures)
The first part of the course will introduce infinite series. Recall Eulerβs identity
often brought up as an example of a beautiful mathematical result. But what exactly is there to prove? Let us try to understand this equation and its connection to Real Analysis.
First of all,
equivalently,
Both formulae involve an infinite process, which in practice can only approximate
What is
This is where the meaning of Eulerβs identity starts to come across. The βinfinite sumβ for
and
Adding up ten terms of each βinfinite sumβ, we obtain approximately
ββ
These infinite expansions, apparently known already to Madhava (ca. 1400), are enormously important in mathematics and have many compelling applications in the sciences. To prove these formulas is the same as to prove the
formula
-
β’ formally define infinite series and make sense of βsums of infinitely many numbersβ;
-
β’ learn about ways to tell whether a given series converges, i.e., has a sum;
-
β’ understand power series which consist of power of
with some coefficients, and see why they define βsmoothβ continuous functions of if they converge.
Expansion of functions as power series is intimately connected with differentiation, a formal treatment of which begins the second part of the course. Higher derivatives of a function are key to approximating the function by polynomials, called Taylor polynomials. It is this theory that allows us to write a βgoodβ function as sum of a Taylor series.
The course concludes with the third part devoted to rigorous treatment of integration. A key result is the Fundamental Theorem of Calculus, which demonstrates that integration is truly a reverse operation to differentiation.
Further study of series: a power series is a sum of infinitely many functions of the form
End of the informal introduction.
The Inverse Function Theorem
We begin with a result which could have been proved in Mathematical Foundations and Analysis (MFA), given more time. The Inverse Function Theorem will be used to define a function
Definition.
Let
increasing, | if | |
strictly increasing, | if | |
decreasing, | if | |
strictly decreasing, | if |
A function satisfying one of the above conditions is called (strictly) monotone.
The above applies to sequences which are functions on
We can use monotone sequences to calculate limits of functions. The following is an MFA-style result:
Lemma 1.1: limit of
For a function
-
(i)
-
(ii) For all strictly increasing sequences
such that as one has
Remark: the Lemma can be expressed in words as follows:
The limit of
at from the left is the common limit of all sequences where a sequence is strictly increasing and converges to
-
Proof of the Lemma (not given in class). (i)
(ii): let be arbitrary. First, we use the definition of to generate such that for allNow we let
be a strictly increasing sequence, and use the above in the definition of β as β to generate such that implies That is, Since the sequence is strictly increasing with limit no term can exceed so in factBut then, by the choice of
for all We have shown that satisfies the definition of limit for the sequence and so (ii) is proved.(ii)
(i): to prove the contrapositive of this implication, we assume that the statement β β is false. This means that there exists some such that for all the interval contains a point, say withChoose
and construct We haveThen, for each
choose and constructSince
by the Sandwich Rule as Also, since the sequence is strictly increasing.We still have, by construction, that
for all Therefore, fails to satisfy the definition of the limit of the sequence i.e. (ii) is false. β‘
Figure 1.1: Lemma 1.1 says that the limit of
The Lemma is illustrated by Figure 1.1. The next result mirrors the Lemma to deal with a limit from the right:
Corollary: limit of
For a function
-
(i)
-
(ii) For all strictly decreasing sequences
such that as one has β‘
In other words,
the limit of
at from the right is the common limit of all sequences where a sequence is strictly decreasing and converges to
We are now ready to prove the first theorem of the course.
Theorem 1.2: the Inverse Function Theorem for strictly increasing functions.
A strictly increasing continuous function
-
Proof.
is surjective, because for every the Intermediate Value Theorem (and its corollary in MFA) gives in such thatA strictly increasing
is injective: indeed, if then either and so or and so In either caseWe have shown that
is bijective, hence it has an inverseWe prove that
is strictly increasing by contradiction. Assume not, then there exist such that and Since is increasing, Since this reads but at the same time a contradiction.We prove that
is continuous at an arbitrary point of its domain by verifying the criterion of continuity, seen in MFA:By results from MFA, this is equivalent to
We first show that
We would like to use Lemma 1.1 for this, so we let be a strictly increasing sequence in which converges to Since is a strictly increasing function, is a strictly increasing sequence; it is also bounded (lies in ), hence by a result from MFA, has a limit, say inThen by Lemma 1.1,
which is as is continuous. Since this says that Thus, henceWe have proved that the common limit of all sequences
where strictly increases and converges to is By Lemma 1.1, this means thatThe proof that
is completely similar, based on the Corollary to Lemma 1.1, and we omit it. Continuity of at is proved. β‘
Example: the
Let
Solution: define
We compose continuous, strictly increasing functions to define a rational power function:
Example: raising to rational power
Define
Remark: one can deduce from the definition of a rational power that
This allows us to formally define arbitrary real powers of
Definition:
If
If
A disadvantage of this definition is that proving the expected properties of powers such as
Infinite series: definition
Definition: infinite series, convergent series, sum.
For real numbers
The
If the sequence of partial sums converges:
Remarks on the definition: (i)
(ii) Any series that is not convergent is said to be a divergent series.
(iii) A series can start from
-
β’
is the th partial sum, -
β’
is the st partial sum, is the nd partial sum, and so on.
Basic examples of convergent/divergent series are discussed in week 1 supervision classes.
Alert: a strict definition of convergence.
The definition of convergence of the series
Weaker definitions can assign a βsumβ to some particular types of series which we consider divergent: CesΓ ro sum, Abel sum etc. They are used in specialist applications which are beyond this course.
The geometric series
The next example is simple yet important: we will see that more complicated series can be studied by comparing them to a geometric series. We revisit a result seen in MFA.
Proposition 1.3: convergence and sum of geometric series.
Let
is convergent if
-
Proof. The
th partial sum of the series is The calculation where the intermediate terms cancel, gives us the formulaIf
we recall from MFA that tends to as so by Algebra of Limits of convergent sequences,The sum of the series, is, by definition, the limit of partial sums if it exists. Hence the sum of the geometric series is
as claimed. β‘
Convergence of series with non-negative terms
Unlike the geometric series, usually there is no nice formula for the
Theorem 1.4: boundedness test for non-negative series.
Let a series
-
(i) the partial sums
are bounded above; -
(ii) the series is convergent.
If (i) and (ii) hold, the sum of the non-negative series is the least upper bound,
-
Proof. The partial sums of a non-negative series form an increasing sequence, because
for all We know from MFA that an increasing sequence of real numbers has a limit iff it is bounded above, and then the limit is the supremum of the terms of the sequence. β‘
Corollary: only two convergence types for non-negative series.
A non-negative series
-
β’ convergent with a non-negative finite sum:
or -
β’ divergent if
In the latter case we use the symbolic notation β
Alert.
Notation
The next test is used very often.
Theorem 1.5: the comparison test for non-negative series.
Assume that
If
To use the Comparison Test, we need to compare with some easy series
Theorem 1.6: the Ratio Test for positive series.
For a positive series
-
Proof. The case
Choose a positive such that For example, works.Since
there exists such that for all Write this as where Then and, repeating this, we obtain We have The upper bound that we have obtained is a finite constant which does not depend on Thus, partial sums of the series are bounded, so by Theorem 1.4, the series is convergent.The case
Put There is such that for But so equivalently for In particular, all for are greater than the positive constant Hence which is unbounded. β‘
Alert: the Ratio test may be inconclusive.
If
In the following test (not taught in lectures, not examinable), we use the
Theorem 1.7: The
For a non-negative series
Remark: Again, if
-
Proof. (not given in class: very similar to the proof of the Ratio Test; not examinable.) The case
Choose a positive so that is still less than for example, works.By definition of limit, there is
such that for all Then for so partial sums of the series are bounded by implying convergence.The case
Put Since is the limit of there is such that for But so equivalently for We therefore have which is unbounded. β‘
Version 2025/02/02 Week 1 in PDF All notes in PDF To other weeks