ISCA明哲论坛：NO.103 Integrating EV Fleet Flexibility into Robust Optimal Planning and Operation of Local Power Grids -- 研究院论坛 -- 现代供应链管理研究院

学术活动

当前位置：首页学术活动研究院论坛 ISCA明哲论坛：NO.103 Integrating EV Fleet Flexibility into Robust Optimal Planning and Operation of Local Power Grids

ISCA明哲论坛：NO.103 Integrating EV Fleet Flexibility into Robust Optimal Planning and Operation of Local Power Grids

2024年07月01日

报告题目： Integrating EV Fleet Flexibility into Robust Optimal Planning and Operation of Local Power Grids

报告人：Gar Goei Loke

报告时间：2024年07月01日（星期一），10:00-11:30

报告地点：明哲楼504

主办单位：东北财经大学现代供应链管理研究院

【报告人简介】

Gar Goei Loke现为杜伦大学商学院管理与营销系副教授。曾担任伊拉斯姆斯大学管理学院助理教授以及新加坡国立大学商学院访问助理教授。他于新加坡国立大学获得博士学位，并在英国剑桥大学获得学士和硕士学位。他的主要研究重点包含不确定条件下的决策，以及通过开发模型、框架、方法和算法，帮助决策者利用数据制定决策等。近期，他的研究重点在于机器学习和优化，并探索如何将两者结合的新方法。他的研究成果主要应用于运营管理、供应链管理、医疗运营管理以及能源和水等业务领域。目前他的研究成果已发表在Operations Research和Manufacturing & Service Operations Management等国际著名期刊上。

【摘要】

The rise of electric vehicles (EVs) poses a challenge to balancing supply and demand in local power grids. In this paper, we examine the implications of EVs and smart charging to both the planning-level question of optimal battery and grid capacity optimization, and the operational-level question of day-to-day supply and demand balancing. We show that an entire EV fleet can be modelled as a battery with stochastically varying capacity, and in doing so, restrict EV charging decisions as a constraint of arriving and departing vehicles. Additionally, there is a specific manner in which the number of arriving and departing EVs are correlated and we propose a novel uncertainty set that relates them via their length-of-stay distribution. Our numerical experiments illustrate that ignoring this specific nature of the correlation, say using the more traditional way of using second-order moments to define a moments uncertainty set, can lead to solutions that fail to be robust, or fail to even solve. We also illustrate that while EVs do lend a degree of flexibility to power grids, a cost needs to be paid either in terms of compelling customers to commit to the time window that their vehicles are parked, or a higher level of uncertainty to the grid operator should there be variability in this time window. The modelling method that we propose in this paper has the potential to be extended to other situations where optimization is required over an inner sub-problem with scheduling decisions on an uncertain number of entities.