PRISTINE TRANSFINITE GRAPHS
AND PERMISSIVE ELECTRICAL NETWORKS
A.H. Zemanian |
PREFACE
Georg Cantor's invention of transfinite numbers occurred
over one hundred years ago and profoundly influenced the development
of twentieth-century mathematics. In fact, it led to a
thorough examination
of the foundations of mathematics. In contrast to the theory
of numbers,
graph theory remained on ``this side of infinity'' until a decade
ago.
An initial embryonic idea concerning ``connections at infinity'' was
introduced in 1975, but it was only after 1987 that transfinite
graphs and
networks were investigated on a continuing basis. This has
enriched the theories of graphs and networks with radically
new constructs and research problems. In general, many solved
problems have now reopened transfinitely and comprise a largely
unexplored
research area. Moreover, there are problems concerning
transfinite graphs
and networks having no counterparts in conventional theories.
Transfinite graphs and networks have already been examined in two
prior books.
``Infinite Electrical Networks'' (Cambridge University Press, 1991)
discussed them at an early stage of development, and then
``Transfiniteness for Graphs, Electrical Networks, and Random
Walks''
(Birkhauser-Boston, 1996) explored them in great generality.
So, why should a third book on this recently conceived subject be
offered?
Simply because much progress has been achieved in the intervening
years in two ways. First, a variety of new results have been
obtained.
Second, a much simpler rendition of the subject has been devised.
This books reports on that continuing research and presents results
beyond
those appearing in those prior books. Those prior works aimed for
generality
and struggled with a variety of difficulties resulting from
the inherent complexity of that subject. A principal objective
of the present
rendition is to provide a much simpler exposition, sacrificing some
generality but capturing the essential ideas of transfiniteness for
graphs and networks. On the other hand, such simplification
enables
the establishment of a variety of new results, such as a
generalization
of Minty's powerful theory for nonlinear monotone networks to
transfinite networks.
One complicating facet of transfinite graphs is that transfinite
nodes can
contain nodes of lower ranks. By assuming away such nodes, we
can
disentangle much of transfinite graph theory. Graphs wherein
no node
contains a node of lower rank will be called ``pristine.''
Actually,
no generality is lost so far as connectedness ideas are concerned
because
nodes can be removed from nodes of higher ranks through ``extraction
paths,''
thereby rendering any transfinite graph into a pristine one.
A second difficulty with the prior exposition of transfinite graphs
and
electrical networks was that transfinite nodes were constructed in a
strictly graph-theoretic manner that completely ignored their
suitability for the flow of electrical currents. With regard
to electrical networks, a transfinite node serves no purpose if
electrical
current cannot flow through it. The presence of such useless
nodes caused
unnecessary trouble. Our present approach avoids such bother
by constructing
transfinite graphs in a special way to ensure that only those
transfinite
nodes that can transmit current need be considered.
Transfinite electrical
networks having such graphs will be called ``permissive.''
This book is organized as follows. After an introductory
chapter,
pristine transfinite graphs are defined in Chapter 2. Our
definitions
of transfinite paths are more concise and much simpler than those in
the
previously mentioned books. Furthermore, ``sections'' and
``subsections''
now coincide, and there is therefore no need to consider
``subsections.''
Also, there no longer are arrow-omega nodes. Such
modifications carry over
to Chapter 3 wherein transfinite graphs are explored. Our
extended discussion
of nu-sequences in Section 3.4 of ``Transfiniteness - for Graphs,
Electrical Networks, and Random Walks,'' used in the
examination of transitivity for $\nu$-connectedness, is now entirely
avoided.
The short but important Section 3.5 in this book defines ``local
finiteness''
for transfinite graphs, a condition we often use. A new result
is presented
in Section 3.6, which extends transfinitely the idea of an ``end''
of a
conventionally infinite graph.
Starting with Chapter 4, we turn to electrical networks. Our
new approach
to transfinite networks is now based on certain metric spaces that
account
for the distribution of resistances throughout the network.
Distances between nodes are measured by metrics, one for each rank,
which pick out those extremities, called ``terminals,'' that
are accessible to electrical currents. ``Permissive
transfinite nodes''
are then constructed by shorting together terminals.
Chapter 5 examines current-voltage regimes in linear networks
based on Tellegen's equation. This was the approach adopted in
previously mentioned two books, but now much simplification is
achieved.
There is no longer any need for basic currents; loop currents
alone will do. Moreover, node voltages always exist and are
unique,
and Kirchhoff's laws are always satisfied. None of this was
true previously.
Chapter 6 presents an entirely new theory for nonlinear transfinite
networks based on Minty's classical theory for finite monotone
networks
and Calvert's generalization to conventionally infinite networks.
The theory is founded on Kirchhoff's laws, no use being made of
Tellegen's equation. Some classical results on finite
nonlinear networks
are needed and presented in Sections 6.2 through 6.4:
Minty's colored-graph theorem, Wolaver's no-gain property, and the
earliest theory of nonlinear networks due to Duffin. The Minty-Calvert
theorem follows in Section 6.5. Then, our new theory of
transfinite
monotone networks appears in Sections 6.6 through 6.10, and this is
related in Section 6.11 to the Tellegen-based theory of Chapter 5.
Chapter 7 takes up maximum principles for node voltages in linear
transfinite networks, and Chapter 8 examines random walks on
such networks. Here, too, much simplification is achieved.
Because transfinite nodes were defined in strictly graph-theoretic
terms in
prior books, some severe assumptions had to be imposed in order to
derive
maximum principles and random walks. A more natural and
concise
approach to these subjects is achieved in this book as a result
of two previously mentioned facts: our permissive
transfinite nodes are amenable to electrical regimes and
the complications arising from nonpristine nodes are now avoided.
One other distinction between ``Transfiniteness - for Graphs,
Electrical
Networks, and Random Walks'' and this book is worth mentioning.
In order to achieve the said maximum principles and transfinite
random walks,
a quite complicated structure, referred to as ``permissively
finitely
structured networks,'' was imposed in the prior book. For the
same purpose,
a simpler structure defined by Conditions 5.3-1 herein is imposed in
this
book. Neither structure subsumes the other. In this way, we
have distinct
results.
Some knowledge of functional analysis is needed for a comprehension
of the analytical parts of this book, but nothing beyond the most
commonly
known facts concerning metric and Hilbert spaces is required.
For the sake of specificity, we refer the reader to some
popular textbooks when citing various standard ideas and theorems,
but there are indeed many other textbooks that can be so used.
Finally, not much is needed from Cantor's theory of
transfinite numbers; if need be, the reader might refer to the
concise
survey in Appendix A of the ``Transfiniteness ---'' book or to any
standard text on set theory.
Scientific books commonly contain mistakes and misprints despite
the best efforts of their authors. Surely, such is the case
for this book, too. As corrections are discovered, they will
be listed
as Errata available on the Internet at www.ece.sunysb.edu/~zeman.
Errata for the prior books, ``Infinite Electrical Networks'' and
``Transfiniteness for Graphs, Electrical Networks, and Random
Walks,''
are also available there.
A.H. Zemanian
Stony Brook, New York |
|