East Coast Optimization Meeting (ECOM) 2019

该奖项为将在弗吉尼亚州费尔法克斯的乔治梅森大学举行的东海岸优化会议(ECOM)提供参与者支持。日期：2019年4月4日-5日。ECOM的目标是向学生和初出茅庐的研究人员介绍当前的优化趋势，并在学术界、工业界和国家实验室之间提供一个强大的网络环境。第一次会议的重点是随机优化。随机优化问题几乎出现在所有的科学和工程领域。随机优化问题的常见例子是：(I)确定金融资产的分配，将受市场变化影响的损失可能性降至最低；(Ii)在第二阶段采油中控制注水井，以最大化地下岩性未知的储油层的净现值；以及(Iii)设计光子超材料，以在不确定的操作环境下最大限度地吸收光。对于这些问题中的每一个，决策者必须在观察不确定的结果(即决策是确定性的)之前选择分配/控制/设计。因此，人们必须适当地量化与每项决定相关的风险，以便控制结果的可变性。这种需求催生了现代随机优化理论，尤其是风险厌恶优化理论。这次会议将为研究生、博士后和其他早期职业科学家提供一个独特的机会，让他们学习两位最优秀的随机优化研究人员的课程，从而帮助培养下一代科学家。此外，该领域的专家将有四次受邀演讲，学生和博士后将有机会通过贡献的演讲来分享他们的工作。会议期间获得的知识将与金融、物理、生物、数据科学、机器(深度学习)和工程等领域相关。该会议由妇女数学协会(AWM)主办。在求解随机优化问题时，准确表达不确定性往往需要大量的样本，这传统上导致了难以解决的非线性优化问题。然而，由于最近在高性能计算、计算模拟和数值优化方面的进展，这类问题的数值解在计算上是可行的。此外，在过去的一年里，四名随机优化研究人员获得了享有盛誉的奖项，其中包括两名丹齐格奖获得者(我们的一位主旨演讲者是其中之一)，一名哈奇扬奖获得者和一名法卡斯奖获得者。因此，对于首届东海岸优化会议来说，随机优化的话题是非常及时的。拟议的会议有可能增进在建模、优化、数值分析、实施和软件开发方面的知识和理解。随机优化涉及统计学、概率论、最优化和变分分析、凸分析和应用数学等多个方面。这次会议将促进这些重要数学领域的新发展。教程和特邀讲座将聚焦于现实生活中的问题，并将讨论处理这些问题的新的优化解算器。因此，与会者可以解决一系列新的具有挑战性的问题。会议上讨论的各种主题，如随机优化、建模、偏微分方程、风险厌恶优化等，都引起了更广泛的兴趣。例如，它们与金融、物理、生物、数据科学和工程相关。预计将有所有这些领域的参与。会议提出的想法将得到积极传播。我们将把讲稿上传到会议网站上。这些资源将有助于创建新的研究生课程。有关会议的更多细节可在http://math.gmu.edu/~hantil/ECOM/2019/.This网站上获得，该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The award provides participant support to the East Coast Optimization Meeting (ECOM) to be held at George Mason University, Fairfax, VA. Date: April 4-5, 2019. The goal of ECOM is to introduce students and early-career researchers to current trends in optimization as well as to provide a strong networking environment between academia, industry, and the national laboratories. The focus of first meeting is Stochastic Optimization. Stochastic optimization problems arise in virtually all science and engineering fields. Common examples of stochastic optimization problems are: (i) determining an allocation of financial assets that minimize the potential for loss subject to market variability; (ii) controlling injection wells in second-stage oil recovery to maximize the net present value of a reservoir in which the subsurface rock properties are unknown; and (iii) designing a photonic meta-material to maximize light absorption subject to uncertain operating environments. For each of these problems, a decision maker must choose an allocation/control/design prior to observing the uncertain outcome (i.e., decisions are deterministic). As a result, one must properly quantify the risks associated with each decision in order to control the outcome variability. This need has led to the modern theory of stochastic and in particular risk-averse optimization. The meeting will provide a unique opportunity for graduate students, postdocs and other early career scientists to take courses from two of the best researchers in stochastic optimization and thus help train next generation of scientists. In addition, there will be four invited talks from the experts in the field and the students and postdocs will have an opportunity to share their work via contributed presentations. The knowledge gained during the meeting will be of relevance to fields such as finance, physics, biology, data science, machine (deep learning), and engineering. The meeting has an affiliation from Association of Women in Mathematics (AWM). Accurately representing the uncertainty when solving stochastic optimization problems often requires an enormous number of samples, which traditionally resulted in intractable nonlinear optimization problem. However, owing to the recent advances in high-performance computing, computational simulation and numerical optimization, the numerical solution of such problems has become computationally feasible. Additionally, this past year, four stochastic optimization researchers received prestigious awards including two Dantzig Award winners (one of our keynote speaker was among the two), a Khachiyan Prize winner and a Farkas Prize winner. For these reasons, the topic of stochastic optimization is very timely for the inaugural East Coast Optimization Meeting. The proposed meeting has the potential to advance knowledge and understanding in modeling, optimization, numerical analysis, implementation and software development. Stochastic optimization encompasses numerous aspects from statistics, probability theory, optimization and variational analysis, convex analysis, and applied mathematics. The meeting will stimulate new developments in these important areas of mathematics. The tutorials and invited talks will focus on real life problems and will discuss new optimization solvers to handle these problems. Thus the attendees can tackle new set of challenging problems. The variety of topics discussed in the meeting, stochastic optimization, modeling, partial differential equations, risk averse optimization is of much wider interest. For instance, these are relevant in finance, physics, biology, data science, and engineering. Participation from all these fields is expected. The ideas created in the meeting will be actively disseminated. We will upload the lecture notes on the conference website. These resources will help create new graduate courses. More details about the meeting are available at http://math.gmu.edu/~hantil/ECOM/2019/.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.