Quantitative arithmetic geometry

点击翻译按钮获取中文摘要

A Diophantine equation is a polynomial equation where one is interested in finding solutions in the whole numbers. Mathematicians have been fascinated by such equations since antiquity. Often simple to state problems require very difficult tools to solve (one of the most famous examples of this being Andrew Wiles's celebrated proof of Fermat's last theorem). Moreover such equations, whilst originally viewed as nothing but a curiosity, have found important applications in modern times to information security and cryptography.Given a Diophantine equation, a fundamental problem is to determine whether a solution actually exists. This problem in itself is very difficult. Things get even more interesting if one has a *family* of Diophantine equations (given by varying the coefficients of the equations, say). In this case one would like to understand the distribution of equations in the family with a solution. This is a very popular modern topic, with Manjul Bhargava being awarded the Fields Medal in 2014 for his work on such problems (this is a kind of mathematician's version of the Nobel prize).The project concerns problems of this type. Here there is a conjecture due to Jean-Pierre Serre, a famous French mathematician, on the distribution of Diophantine equations in certain families with a solution (namely plane conics). We will answer some cases of Serre's problem, as well as extending Serre's original framework to more general problems.A famous theorem of Erdos and Kac also states that a "random" integer n has approximately log log n prime factors (in a precise probabilistic sense). We will obtain analogues of this probabilistic result in the setting of families of Diophantine equations, where we ask for the number of primes p for which a given equation is not soluble modulo p.