Random matrices, arithmetic combinatorics, and incidence geometry

随机矩阵、算术组合和关联几何

• 批准号: 1266164
• 负责人: Terence Tao
• 金额: $ 75万
• 依托单位: University of California-Los Angeles
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2013
• 资助国家: 美国
• 起止时间: 2013-07-01 至 2019-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1266164&HistoricalAwards=false
• 关键词: Random; matrices; arithmetic; combinatorics; incidence

This proposal seeks to make progress in three distinct (and only looselyrelated) areas in mathematics. Firstly, in the area of random matrix theory, we seek to build upon recent progress in understanding the fascinating universality phenomenon for such matrices, by generalising the ensembles for which universality can be established, and looking for the underlying mechanisms causing this universality. Secondly, in the area of arithmetic combinatorics, we seek to improve our theory of approximate groups and related structures in noncommutative settings, with potential impact on geometric group theory and finite group theory. Finally, we wish to extend and understand the exciting new connections between algebraic geometry, algebraic topology, and incidence combinatorics which have led to remarkable breakthroughs in the latter subject by applying algebraic methods.Physical systems are often incredibly complicated, being driven by interactions between countless components of those systems. However, in many cases, a remarkable phenomenon known as universality occurs, in which the macroscopic statistics of the a wide variety of systems are governed by a single universal mathematical law, almost without any regard to how the components of these systems interact with each other at the microscopic level. For instance, the famous central limit theorem in mathematics asserts that many statistics (e.g. the distribution of heights in a human population) are described by a single curve known as the "bell curve" or "Gaussian distribution". While the central limit theorem is now very well understood, there are other universal laws that are still mysterious, such as the Dyson sine law that has been empirically found to govern such diverse statistics as nuclear scattering, arrival times of buses, and the spacing of the zeta function in number theory. However, we have recently begun to understand this law for a simple class of models known as random matrix models, though even for these models there is still much work to be done. We plan to work on further understanding the underlying causes of universality for random matrix models and related models, with the hope of shedding insight on the universality phenomenon for other models as well.The second aspect of the research program concerns the phenomenon of expansion in various networks, most famously manifested by the "six degrees of separation" experiment of Milgram, that asserts that any two people in the world are linked by at most six degrees of acquaintance. For a mathematical model of this phenomenon known as a Cayley graph, the existence of expansion is closely tied to the absence of a certain type of mathematical object known as an approximate group. There has been a substantial amount of progress in understanding what exactly approximate groups look like, but the control we have on these groups is still not as precise as we would like. We plan to study the theory of approximate groups further, with an eye towards applications to other areas of mathematics such as understanding the geometry of groups.Combinatorial incidence geometry is the study of how simple geometric objects such as lines, circles, and points can be arranged together in as efficient a configuration as possible for various purposes (e.g. to try to arrange a given number of points and lines so that as many points lie on as many lines as possible). Understanding the limits of how efficiently one can design such configurations is a basic mathematical question which also has practical applications (e.g. in designing frequency configurations for cell phones in order to maximize capacity and minimize interference). Recently, there has been several breakthroughs in the subject by introducing methods from algebra (particularly algebraic geometry and algebraic topology). This unexpected development is still not well understood; the algebraic methods can solve some problems almost completely, while making no mark on other ostensibly similar problems. We plan to study these methods further and understand exactly what their strengths and limitations are.

这一建议试图在数学的三个不同（且仅松散相关）的领域取得进展。 首先，在该地区的随机矩阵理论，我们试图建立在最近的进展，了解这种矩阵的迷人的普遍性现象，通过概括的合奏，可以建立普遍性，并寻找导致这种普遍性的潜在机制。 其次，在算术组合学领域，我们寻求改进我们的近似群理论和非交换设置中的相关结构，对几何群论和有限群论产生潜在影响。最后，我们希望扩展和理解代数几何，代数拓扑和关联组合学之间令人兴奋的新联系，这些联系通过应用代数方法在后一个主题中取得了显着的突破。物理系统通常非常复杂，由这些系统的无数组件之间的相互作用驱动。 然而，在许多情况下，一个被称为普适性的显着现象发生了，其中各种各样的系统的宏观统计数据由一个普遍的数学定律支配，几乎没有考虑这些系统的组件如何在微观水平上相互作用。 例如，数学中著名的中心极限定理断言，许多统计数据（例如人口中的身高分布）可以用一条称为“钟形曲线”或“高斯分布”的曲线来描述。 虽然中心极限定理现在已经被很好地理解了，但还有其他一些普遍的定律仍然是神秘的，例如戴森正弦定律，它被经验地发现用于管理核散射，公共汽车到达时间和数论中zeta函数的间距等各种统计数据。 然而，我们最近开始了解一类简单模型（称为随机矩阵模型）的这一定律，尽管即使对于这些模型，仍然有很多工作要做。 我们计划进一步了解随机矩阵模型和相关模型普适性的根本原因，希望对其他模型的普适性现象也有深入的了解。研究计划的第二个方面涉及各种网络中的膨胀现象，最著名的是Milgram的“六度分离”实验，它断言世界上任何两个人最多有六个熟人度的联系。 对于这种现象的数学模型称为凯莱图，扩展的存在与缺乏某种类型的数学对象称为近似群密切相关。 在理解近似群的确切样子方面已经取得了相当大的进展，但是我们对这些群的控制仍然没有我们想要的那么精确。 我们计划进一步研究近似群的理论，着眼于应用到数学的其他领域，如理解群的几何。组合关联几何是研究如何简单的几何对象，如直线，圆，并且点可以为了各种目的而以尽可能有效的配置布置在一起（例如，试图安排给定数量的点和线，以便尽可能多的点位于尽可能多的线上）。 了解如何有效地设计这种配置的限制是一个基本的数学问题，也有实际应用（例如，在设计蜂窝电话的频率配置，以最大限度地提高容量和最小化干扰）。 近年来，通过引入代数（特别是代数几何和代数拓扑）的方法，在这一问题上取得了一些突破。 这个意想不到的发展仍然没有得到很好的理解;代数方法可以解决一些问题几乎完全，而没有标记在其他表面上类似的问题。 我们计划进一步研究这些方法，并确切了解它们的优势和局限性。