Edinburgh Mathematics Programme
Level 2/3 Module
MS0105 (FoA) Fundamentals of Analysis (S2)
Introduction
The basic tools of analysis which you will have seen in the FoC (Foundations of Calculus) module are developed in this course by introducing powerful concepts to enable us to solve deep problems of Analysis. One of the most important concepts
covered in this module is that of completeness, which arises in
many guises. The natural language for discussing this and other
important notions, such as convergence, is that of metric
spaces. Many results of analysis depend on studying spaces of functions and ways of measuring functions in these
spaces. Once again, the language of metric spaces is ideal. This module will introduce these concepts by first understanding how they work in R and R^n, then passing to the metric space setting, with
particular emphasis on metric spaces of functions.
Aims
- To cover basic ideas of convergence and completeness in R^n.
- To discuss the basic notions of topology in R^n.
- To introduce metric spaces and the notions of convergence in a
metric space and completeness of a metric space.
- To introduce the idea of uniform convergence.
- To introduce the basic concepts of function spaces as metric
spaces, with illustrations.
Prerequisites
MS0095 FoC; concurrent or previous attendance at MS0225 LAG; concurrent or
previous attendance at MS0226 MAM would be helpful.
Syllabus summary
Properties of the reals. Cauchy sequences; completeness of R^n. Contraction mapping theorem in R with basic applications. Open
and closed subsets of R^n. Definition and examples of metrics.
Convergent
and Cauchy sequences in metric spaces. The uniform metric
on C[a,b] and the notion of uniform convergence. Completeness of C[a,b]
with the uniform metric. Examples of other function spaces as metric
spaces.
Text
- Stoll, M., Introduction to Real Analysis, Addison-Wesley, 1997 (2nd Ed., 2001)
Library shelf reference: QA 300 Sto
Syllabus
- Definition of supremum and infimum, basic properties, statement of LUB axiom; consequences: Archimedean property, density of rationals, etc. (2)
- The completeness of R. Subsequences, the nested interval and Bolzano-Weierstrass theorems in R. Cauchy sequences, convergent implies Cauchy, Cauchy in R implies convergent. (2.5)
- Contraction Mapping Theorem for mappings on closed intervals in R;
application to finding approximate roots by iteration. (1)
- Sequences in C, convergence equivalent to convergence of real and
imaginary parts. The Bolzano-Weierstrass theorem and Cauchy sequences in C.
The completeness of C.
Inner product and euclidean norm on R^n. Cauchy-Schwarz and triangle
inequalities. Convergent and Cauchy sequences in R^n. The completeness
of R^n,
the Bolzano-Weierstrass in R^n. (2.5)
- Infinite series with complex terms. The Cauchy criterion for
convergence of infinite series.
Absolute convergence implies convergence. Complex power series. Radius
and disc of convergence. (2.5)
- Definition of open and closed sets in R^n; basics: intersection and union properties, etc., closed if and only if every limit point lies in the set. Continuity for a function f: A --> R with A a subset of R^n, attainment of bounds when A is closed and bounded. (1.5)
- Definition of a metric, examples of different metrics on R^n and
on C[a,b]. Convergence in metric spaces, Cauchy sequences and completeness. Continuity in metric spaces. Open and closed sets in metric spaces. (3.5)
- Pointwise and uniform convergence. Cauchy condition for uniform convergence. Preservation of continuity under uniform convergence (hence C[a,b] complete under uniform metric). Uniform convergence of series of functions, the Weierstrass M-test. (3.5)
- The convergence of Fourier series (pointwise/uniform), uniform
convergence of Fejer means for continuous functions, convergence in the mean-square metric d_2 of Fourier
series of continuous functions and Parseval's Theorem; Weierstrass approximation theorem. (3)
Notes and Links
- The principal themes of the module are completeness in its many guises
and the ramifications of completeness.
- Sections 1 and 2 of the syllabus are meant to be a deeper and fully rigorous treatment of some of the material covered in FoC.
- This is the first explicit exposure to complex series, so some care needs to be taken over their treatment.
- In this module, metrics provide a language for discussing convergence and
completeness, and so the emphasis is on the definition of a metric space
and giving plenty of useful examples of metrics. It is not
intended that there should be a detailed development of the theory of metric
spaces per se.
- The examples of metrics to be given should include those corresponding
to p=1,2 and infinity, both on euclidean space and on C[a,b]. For p=2 it
is useful to make the link with inner product spaces and the Cauchy-Schwarz
inequality on inner product spaces. The standard inner product,
Cauchy-Schwarz and triangle inequalities on R^n are covered in LAG, but
more general inner products are not.
- Section 9 introduces the idea that the language of metrics is the appropriate one for concrete problems in
analysis such as that of convergence of Fourier series. (The lecturer may wish to tailor the topics suggested here to reflect time remaining and stamina of the class.) In connection with Parseval's theorem and convergence of Fourier
series in the d_2 metric, LAG does not currently cover the "closest
point to the span of finitely many orthonormal vectors" material. The pointwise convergence of Fourier series may
be touched on in MAM (even if not officially in the syllabus for MAM).
Suggest to liaise with MAM lecturer on this topic.
- An alternative way to round off the course would be to replace the Fourier series material in Section 9 with a discussion of the Contraction Mapping Theorem in metric spaces and give applications of this (e.g. Picard's theorem on the solution of ODEs, the inverse mapping theorem). However, there would be a knock-on effect for AaI if this were done.
Outcomes
- Understanding the definitions and concepts of convergence, Cauchy
sequences, and completeness in R and R^n.
- Understanding the notions of open and closed subsets of R^n and
their relationship to convergence.
- Familiarity with the notion of a metric and examples of metrics.
- Familiarity with the notion of convergence in a metric space and
completeness of a metric space.
- Ability to test sequences and series of functions for uniform convergence.
- Familiarity with the uses of the Weierstrass M-test.
- Understanding the notion of function spaces as metric spaces and familiarity
with particular examples: C([a,b]) with d_\infty, d_2 and d_1.
- Familiarity with the idea that problems of analysis such as
convergence of Fourier series can be couched in and resolved using the
language of metric spaces.
TAG/AOD
Last revised 03.03.02