Topics in disordered systems

From MaRDI portal





This small book contains the lecture notes of a post-graduate course on disordered systems given at the Mathematics Department of the ETH-Zürich. They deal with the equilibrium statistical mechanics of Ising-spin systems on lattices with random interactions from the point of view of mathematical physics and probability theory. There are no textbooks and few reviews in this field and so the appearance of this introductory text is most timely and welcome. Rather than giving an overview of the entire field of disordered systems, this booklet concentrates on a narrow class of models, namely Ising-spin systems with random nearest neighbor interactions. These comprise the maybe most interesting model of all, the Edwards-Anderson (EA) model of a spin glass, and, in fact, most of the most interesting questions arising in the theory of disordered systems can be posed (if not answered) in this context. Clearly, the entire text is geared towards the big problem of the field: How to describe and analyse the low-temperature phase of the EA-model? Of the four sections of the book, the first three deal with related, but simpler questions. They are devoted to ground states properties. Section 1 treats the ground states in a model with random, but ferromagnetic (non-negative) interactions. Although two obvious translation invariant ground states are immediately identified, other ground states, containing interfaces may or may not exist, and this problem turns out to be already a very complex and tricky one. Allowing couplings to take arbitrary signs makes the problem even harder. In Section 2 a very peculiar model of that type is considered, in which the strength of certain couplings grows without limit as the volume increases which allows to construct infinitely many ground states in certain situations. Section 3 takes us to the high temperature regime. The issue here is to prove uniqueness of the Gibbs state in a region as large as possible. In all three sections percolation theory is a recurring theme that enters as a main tool to tackle the problems at hand. Section 4 is the largest and most important section of the book. It deals with the central issue of how the low-temperature phase of the EA model could look like, and in particular, to what extent the predictions of mean-field theory (the ``Parisi picture) could be applicable. To investigate this issue, a rather new and original conceptual approach to the mathematical description of disordered systems in terms of so-called ``meta-states (probability measures on the space of Gibbs measures) is introduced and its consequences for the problem at hand are discussed. Some of the more technical details of this approach are elaborated in an appendix. In the opinion of the reviewer, this approach is an important contribution to the field and it is very nice to have an introductory exposition in these notes. This book will be useful for the active researcher in the field but is also accessible to advanced graduate students with some background in statistical mechanics.




Cited in
(44)








This page was built for publication: Topics in disordered systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4380378)