Table of contents

In this paper the topological theory of quasi-periodic functions on the plane is presented. The development of this theory was started (in another terminology) by the Moscow topology group in the early 1980s, motivated by needs of solid state physics which led to the necessity of investigating a special (non-generic) case of Hamiltonian foliations on Fermi surfaces with a multivalued Hamiltonian function [1]. These foliations turned out to have unexpected topological properties, discovered in the 1980s ([2], [3]) and 1990s ([4]-[6]), which led finally to non-trivial physical conclusions ([7], [8]) by considering the so-called geometric strong magnetic field limit [9]. A reformulation of the problem in terms of quasi-periodic functions and an extension to higher dimensions in 1999 [10] produced a new and fruitful approach. One can say that for monocrystalline normal metals in a magnetic field the semiclassical trajectories of electrons in the quasi-momentum space are exactly the level curves of a quasi-periodic function with three quasi-periods which is the restriction of the dispersion relation to the plane orthogonal to the magnetic field. The general study of topological properties of level curves for quasi-periodic functions on the plane with arbitrarily many quasi-periods began in 1999 when some new ideas were formulated in the case of four quasi-periods [10]. The last section of this paper contains a complete proof of these results based on the technique developed in [11] and [12]. Some new physical applications of the general problem were found recently [13].

This survey treats two chapters in the theory of log minimal models, namely, the chapter on different notions of models in this theory and the chapter on birational flips, that is, log flips, mainly in dimension 3. Our treatment is based on ideas and results of the second author: his paper on log flips (and also on material from the University of Utah workshop) for the first chapter, and his paper on prelimiting flips (together with surveys of these results by Corti and Iskovskikh) for the second chapter, where a complete proof of the existence of log flips in dimension 3 is given. At present, this proof is the simplest one, and the authors hope that it can be understood by a broad circle of mathematicians.

Each convex smooth curve on the plane has at least four points at which the curvature of the curve has local extrema. If the curve is generic, then it has an equidistant curve with at least four cusps. Using the language of contact topology, V.I. Arnol'd formulated conjectures generalizing these classical results to co-oriented fronts on the plane, namely, the four-vertex conjecture and the four-cusp conjecture. In the present paper these conjectures and some related results are proved. Along with a simple generalization of the Sturm-Hurwitz theory, the main ingredient of the proof is a theory of pseudo-involutions which is constructed in the paper. This theory describes the combinatorial structure of fronts on a cylinder. Also discussed is the relationship between the theory of pseudo-involutions and bifurcations of Morse complexes in one-parameter families.