Abstract

In the era of dynamic energy grid with price-sensitive demands and intermittent renewable energy sources, it is crucial to optimize the operational cost and manage the operation decisions of energy systems efficiently, while maintaining their states within the stable operating limits. There are two key challenges facing in future energy systems for the sustainable society.

First, integrating diverse energy systems into today’s dominant alternating current (AC) electric power systems creates an increasingly hard optimization problem. The hardness stems from two issues: (1) non-convexity involving complex-valued entities of AC electric power systems, and (2) combinatoric constraints involving discrete control decisions. There have been recent advances in applying convex relaxations to tackle non-convex AC electric power systems. On the other hand, typical combinatorial optimization studies consider linear constraints, instead of the non-linear constraints of AC electric power systems.

Second, while it was feasible to rely on static offline planning based on historic data in the past, today’s energy systems are facing drastic uncertainty and volatility in the presence of fluctuating electricity markets and renewable energy supplies. Effective online decision-making mechanisms are required to enable real-time operation decisions and capacity provisioning in energy systems. It is vital to provide good performance guarantee to energy system management, without asymptotic or specific stochastic assumptions, while allowing integration of data-driven prediction information.

This talk presents a comprehensive study on the optimization and management of sustainable energy systems. It covers combinatorial optimization of AC electric power systems with (inelastic) discrete demands, with efficient approximation algorithms and inapproximabilty results. Also, it covers competitive online algorithms for energy system management, with energy storage and distributed energy generations, and applications to datacenters. This talk is partially based on two monographs to be published by Now Publishers and Morgan & Claypool Publishers.

Speaker's Biography

Sid Chi­-Kin Chau is an assistant professor with the Masdar Institute in Abu Dhabi, UAE, which was established in collaboration with MIT. His primary research area is Internet-of-Things and cyber-physical systems for smart cities and smart energy systems, including smart grid, smart buildings, intelligent vehicles and transportation. He also researches in broad areas of wireless communications, algorithms, and big data analytics. Previously, He was a visiting professor with MIT, and a senior research fellow with A*STAR in Singapore. He received the Ph.D. from University of Cambridge and B.Eng. from the Chinese University of Hong Kong. He is a TPC co-chair of ACM e-Energy 2018. He also chairs Special Interest Group in Smart Metering, Demand Response and Dynamic Pricing for IEEE ComSoc Emerging Technical Committee of Smart Grid Communications. Further information about his research can be found at http://www.SmartSustainability.org/