Department of Mathematics and Statistics 620-629 Integrable Models
Semester 2, 2009


Let me first tell you what 620-629 is not about.

Integrable Models (in the sense discussed in 620-629) have very little to do with modeling physical phenomena (in the sense of courses on continuum mechanics, for example). The subject may have started historically that way (by describing concrete physical phenomena, such as solitary waves in shallow water channels) but since then the emphasis has steadily drifted away from these historical origins.

Furthermore, the integrable nonlinear PDE's that are studied in 620-629 are so very special that the lessons that we learn from solving them do not extend in any useful way to more general nonlinear PDE's. The latter definitely require perturbative methods of different types (regular and singular), which are the subject of entirely different studies.

Now, let me try to give you a rough idea of what the course is about.

As mentioned above, it is a fact of life that almost all PDE's cannot be solved exactly. Definitely, almost all nonlinear PDE's cannot be solved exactly. Like it or not, perturbative techniques remain the one and only way to say something meaningful about general PDE's.

However, there are some very, very special classes of nonlinear PDE's that, amazingly enough, can be solved exactly. These nonlinear PDE's are extremely special and not at all representative of general PDE's that one meets in applied mathematics, mathematical physics, modeling physical phenomena, etc. So, why should we study them?

There are two diagonally opposite answers to this question:

1. One can say that these PDE's and the phenomena that they describe are too special to be worth studying, and that we should turn our attention to general PDE's, learn all that we can about perturbative methods, etc.

2. One can also say that these PDE's and the phenomena that they describe are so special that we definitely need to learn as much as we can about them.

While the first answer is a perfectly valid one from the viewpoint of an honest applied mathematician, in 620-629, we adopt the second. There are two reasons for that.

The first is that, once we decide to fully understand how to solve these very special nonlinear PDE's, a very rich panorama of mathematics opens up to us: Representation theory, complex analysis, algebraic combinatorics, algebraic geometry, symplectic geometry, and more, all come together as tools that are necessary to fully understand what is going on.

My point is that the mathematical experience that we gain from a serious study of Integrable Models makes this study very well worth it.

The second reason is that, it is a fact that the very basic ideas of Classical (and Quantum) Integrable Models, such as tau functions, vertex operators, infinite dimensional hierarchies, etc, etc, have recently (and totally unexpectedly) appeared in modern mathematical physics (Seiberg-Witten theory, low dimensional string theory, topological string theory, etc) and in pure mathematics (the study moduli spaces of Riemann surfaces, Geometric Langlands, etc).

To mathematical physicists (of the more mathematical type), these are very interesting developments with intriguing inter-connections that beg to be fully understood.

Back to homepage
top of page