To save content items to your account,
please confirm that you agree to abide by our usage policies.
If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account.
Find out more about saving content to .
To save content items to your Kindle, first ensure no-reply@cambridge.org
is added to your Approved Personal Document E-mail List under your Personal Document Settings
on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part
of your Kindle email address below.
Find out more about saving to your Kindle.
Note you can select to save to either the @free.kindle.com or @kindle.com variations.
‘@free.kindle.com’ emails are free but can only be saved to your device when it is connected to wi-fi.
‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.
The 400-year-old Kepler conjecture asserts that no packing of congruent balls in three dimensions can have a density exceeding the familiar pyramid-shaped cannonball arrangement. In this book, a new proof of the conjecture is presented that makes it accessible for the first time to a broad mathematical audience. The book also presents solutions to other previously unresolved conjectures in discrete geometry, including the strong dodecahedral conjecture on the smallest surface area of a Voronoi cell in a sphere packing. This book is also currently being used as a blueprint for a large-scale formal proof project, which aims to check every logical inference of the proof of the Kepler conjecture by computer. This is an indispensable resource for those who want to be brought up to date with research on the Kepler conjecture.
The discrete cube {0, 1}d is a fundamental combinatorial structure. A subcube of {0, 1}d is a subset of 2k of its points formed by fixing k coordinates and allowing the remaining d − k to vary freely. This paper is concerned with patterns of intersections among subcubes of the discrete cube. Two sample questions along these lines are as follows: given a family of subcubes in which no r + 1 of them have non-empty intersection, how many pairwise intersections can we have? How many subcubes can we have if among them there are no k which have non-empty intersection and no l which are pairwise disjoint?
These questions are naturally expressed using intersection graphs. The intersection graph of a family of sets has one vertex for each set in the family with two vertices being adjacent if the corresponding subsets intersect. Let $\I(n,d)$ be the set of all n vertex graphs which can be represented as the intersection graphs of subcubes in {0, 1}d. With this notation our first question above asks for the largest number of edges in a Kr+1-free graph in $\I(n,d)$. As such it is a Turán-type problem. We answer this question asymptotically for some ranges of r and d. More precisely we show that if $(k+1)2^{\lfloor\frac{d}{k+1}\rfloor}<n\leq k2^{\lfloor\frac{d}{k}\rfloor}$ for some integer k ≥ 2 then the maximum edge density is $\bigl(1-\frac{1}{k}-o(1)\bigr)$ provided that n is not too close to the lower limit of the range.
The second question can be thought of as a Ramsey-type problem. The maximum such n can be defined in the same way as the usual Ramsey number but only considering graphs which are in $\I(n,d)$. We give bounds for this maximum n mainly concentrating on the case that l is fixed, and make some comparisons with the usual Ramsey number.
The study of the phase transition of random graph processes, and recently in particular Achlioptas processes, has attracted much attention. Achlioptas, D'Souza and Spencer (Science, 2009) gave strong numerical evidence that a variety of edge-selection rules in Achlioptas processes exhibit a discontinuous phase transition. However, Riordan and Warnke (Science, 2011) recently showed that all these processes have a continuous phase transition.
In this work we prove discontinuous phase transitions for three random graph processes: all three start with the empty graph on n vertices and, depending on the process, we connect in every step (i) one vertex chosen randomly from all vertices and one chosen randomly from a restricted set of vertices, (ii) two components chosen randomly from the set of all components, or (iii) a randomly chosen vertex and a randomly chosen component.
Böttcher, Schacht and Taraz (Math. Ann., 2009) gave a condition on the minimum degree of a graph G on n vertices that ensures G contains every r-chromatic graph H on n vertices of bounded degree and of bandwidth o(n), thereby proving a conjecture of Bollobás and Komlós (Combin. Probab. Comput., 1999). We strengthen this result in the case when H is bipartite. Indeed, we give an essentially best-possible condition on the degree sequence of a graph G on n vertices that forces G to contain every bipartite graph H on n vertices of bounded degree and of bandwidth o(n). This also implies an Ore-type result. In fact, we prove a much stronger result where the condition on G is relaxed to a certain robust expansion property. Our result also confirms the bipartite case of a conjecture of Balogh, Kostochka and Treglown concerning the degree sequence of a graph which forces a perfect H-packing.
In automata theory, quantum computation has been widely examined for finite statemachines, known as quantum finite automata (QFAs), and less attention has been given toQFAs augmented with counters or stacks. In this paper, we focus on such generalizations ofQFAs where the input head operates in one-way or realtime mode, and present some newresults regarding their superiority over their classical counterparts. Our first result isabout the nondeterministic acceptance mode: Each quantum model architecturallyintermediate between realtime finite state automaton and one-way pushdown automaton(one-way finite automaton, realtime and one-way finite automata with one-counter, andrealtime pushdown automaton) is superior to its classical counterpart. The second andthird results are about bounded error language recognition: for anyk > 0, QFAs with k blind counters outperform theirdeterministic counterparts; and, a one-way QFA with a single head recognizes an infinitefamily of languages, which can be recognized by one-way probabilistic finite automata withat least two heads. Lastly, we compare the nondeterminictic and deterministic acceptancemodes for classical finite automata with k blind counter(s), and we showthat for any k > 0, the nondeterministic models outperform thedeterministic ones.
In this paper, we define k-counting automata as recognizers forω-languages, i.e. languages of infinite words. Weprove that the class of ω-languages they recognize is a proper extensionof the ω-regular languages. In addition we prove that languagesrecognized by k-counting automata are closed under Boolean operations. Itremains an open problem whether or not emptiness is decidable fork-counting automata. However, we conjecture strongly that it is decidableand give formal reasons why we believe so.
It is well known that a graph with m edges can be made triangle-free by removing (slightly less than) m/2 edges. On the other hand, there are many classes of graphs which are hard to make triangle-free, in the sense that it is necessary to remove roughly m/2 edges in order to eliminate all triangles.
We prove that dense graphs that are hard to make triangle-free have a large packing of pairwise edge-disjoint triangles. In particular, they have more than m(1/4+cβ) pairwise edge-disjoint triangles where β is the density of the graph and c ≥ is an absolute constant. This improves upon a previous m(1/4−o(1)) bound which follows from the asymptotic validity of Tuza's conjecture for dense graphs. We conjecture that such graphs have an asymptotically optimal triangle packing of size m(1/3−o(1)).
We extend our result from triangles to larger cliques and odd cycles.
We prove that the threshold for the appearance of a k-regular subgraph in Gn,p is at most the threshold for the appearance of a non-empty (k+1)-core. This improves a result of Pralat, Verstraete and Wormald [5] and proves a conjecture of Bollobás, Kim and Verstraete [3].
In this paper we analyse classical Maker–Breaker games played on the edge set of a sparse random board G ~ n,p. We consider the Hamiltonicity game, the perfect matching game and the k-connectivity game. We prove that for p(n) ≥ polylog(n)/n the board G ~ n,p is typically such that Maker can win these games asymptotically as fast as possible, i.e., within n+o(n), n/2+o(n) and kn/2+o(n) moves respectively.
Summary. This part of this book, which is the first of the four foundational chapters, presents a systematic development of trigonometry, volume, hypermap, and fan. There is a separate chapter on each of these topics. The purpose of the this material is to build a bridge between the foundations of mathematics, as presented in formal theorem proving systems such as HOL Light, and the solution to the packing problem.
In this chapter, trigonometry is developed analytically. Basic trigonometric functions are defined by their power series representations, and calculus of a single real variable is used to develop the basic properties of these functions. Basic vector geometry is presented.
Background Knowledge
formal proof
We repeat that our purpose is to give a blueprint of the formal proof of Kepler's conjecture that no packing of congruent balls in three-dimensional Euclidean space has density greater than the familiar cannonball packing. The blueprint of a formal proof is not the same as a formal proof, which is a fleeting pattern of bits in a computer. The book describes to the reader how to construct the computer code that produces and then reliably reproduces that pattern of bits.
A more traditional book might take as its starting point the imagined mathematical background of a typical reader. The blueprint of a formal proof starts instead with the current mathematical background of a formal proof assistant. I surveyed the knowledge base of my formal proof assistant and compared it with what is needed in the construction of our formal proof. It turns out that the proof assistant already has an adequate background in real analysis, basic topology, and plane trigonometry, including the trigonometric addition laws, and formulas for derivatives.
This book is just a blueprint, which gives instructions about how to construct the formal proof. Flyspeck is the name of an ongoing project to construct a formal proof of the Kepler conjecture in the HOL Light proof assistant, along the lines described in this book. The eventual aim of the project is to give a formal verification of the computer portions of the proof as well as the standard text portions of the proof. The project is about 80% complete as of May 2012. The source code for the project and information about the current project status are available at [21].
Here is the fine print about the current project status. (I hope that this status report is out of date by the time this book is printed.) There are four components to the formalization project (text, hypermaps, linear programs, and nonlinear inequalities), at various stages of completion.
This section gives a brief history of the study of dense sphere packings. Further details appear at [43] and [20]. The early history of sphere packings is concerned with the face-centered cubic (FCC) packing, a familiar pyramid arrangement of congruent balls used to stack cannonballs at war memorials and oranges at fruit stands (Figure 1.1).
Sanskrit sources
The study of the mathematical properties of the FCC packing can be traced to a Sanskrit work (the Āryabhaṭīya of Āryabhaṭa) composed around 499 CE. The following passage gives the formula for the number of balls in a pyramid pile with triangular base as a function of the number of balls along an edge of the pyramid [40].
For a series [lit. “heap”] with a common difference and first term of 1, the product of three [terms successively] increased by 1 from the total, or else the cube of [the total] plus 1 diminished by [its] root, divided by 6, is the total of the pile [lit. “solid heap”].
In modern notation, the passage gives two formulas for the number of balls in a pyramid with n balls along an edge (Figure 1.2):
Harriot and Kepler
The modern mathematical study of spheres and their close packings can be traced to Harriot. His work – unpublished, unedited, and largely undated – shows a preoccupation with sphere packings. He seems to have first taken an interest in packings at the prompting of Sir Walter Raleigh.
Summary. A planar graph, which is a graph that admits a planar embedding, has too little structure for our purposes because it does not specify a particular embedding. A plane graph carries a fixed embedding, which gives it a topological structure where combinatorics alone should suffice. A hypermap gives just the right amount of structure. It is a purely combinatorial object, but carries information that the planar graph lacks by encoding the relations among nodes, edges, and faces. This chapter is about hypermaps.
In the original proof of the Kepler conjecture, the basic combinatorial structure was that of a planar map, as defined by Tutte [47]. Although planar maps appear throughout that proof, they are lightweight objects, in the sense that no significant structural results are needed about them.
Gonthier makes hypermaps the fundamental combinatorial structure in his formal proof of the four-color theorem [16]. His formal proof eliminates topological arguments such as the Jordan curve theorem in favor of purely combinatorial arguments. When I learned of Gonthier's work, I significantly reorganized the proof by replacing planar maps with hypermaps, making them heavyweight objects, in the sense that significant structural results about them are needed.
As a result of these changes, many parts of the proof that were originally done topologically can now be done combinatorially, a change that significantly reduces the effort required to formalize the proof. These changes also make it possible to treat rigorously what was earlier done by geometric intuition.