Latin trades and critical sets in latin squares
Nicholas J. Cavenagh · The University of Queensland · 2003
Not cached yet. This record was fetched live for this request and is stored after the lookup. Reopening the page should serve the cached copy.
The University of Queensland · The University of Queensland · dissertation · 2003
Abstract
Each of the problems presented in this thesis is related in some way to the following question: Given an integer n, for what values m do there exist critical sets of order n? The methods of analysis/attack of this problem include: • counting arguments, • the development of latin trade theory, and • spectrum constructions, exploiting the multitude of ways in which a latin square can be defined and represented. In Chapter 2 we approach critical sets in a roundabout way by analysing the properties of their complements. We develop a theory of constructing critical sets by building the complements first, with a view to obtaining results on weak critical sets.
The properties of latin trades in the back circulant square are studied in Chapter 4. Our motivation is that the underlying group structure can be exploited to obtain results on latin trades. Moreover latin trades in the back circulant latin square are shown to exhibit geometric properties, which may be fundamental to latin trade theory. Some of our constructions are used later in Chapter 7 to obtain spectrum results on critical sets in back circulant latin squares. In Chapter 5 we take another approach and locate latin trades in arbitrary sets of three rows in latin squares. This yields a result in Chapter 6 on the size of the smallest critical sets when there is an empty row.
In this chapter we also give a counting argument which gives a bound on the size of the largest 2-critical set in a latin square. In Chapter 8 we construct a class of critical sets in a cross product of back circulant latin squares, by modifying an existing class of critical sets. While the latin trade constructions from Chapter 4 make the proof manageable, the result indicates that that modifying known critical sets to make new ones may be impractically complicated in some instances. A special type of critical set with balanced properties is introduced in Chapter 9. Latin trades are again used to prove the criticality of these sets.
A representation of a latin square as a colouring of a complete bipartite graph is given in Chapter 10 which is used to prove the existence of small critical sets in arbitrary latin squares.
Open the resolved locator · 10.14264/106295
Metadata licence: OpenAlex metadata (CC0)