The design, upgrade, and real-time operation of a power network are all contingent upon sev- eral optimization problems that are hard to solve due to two main reasons: (i) the nonlinearity induced by the laws of physics, and (ii) the presence of integer variables. The existing solvers for energy-related optimization problems either make potentially conservative approximations or deploy general-purpose local-search algorithms to handle the non-convexities, which may incur tens of billions of dollars annually. The general intractability of power optimization problems has a direct impact in the practice of energy efficiency, and this is arguably one of the most fundamen- tal issues that hold back the power engineering. The primary objective of this project is to address this non-convexity issue by developing high-performance optimization techniques that can be ap- plied to a broad set of nonlinear energy problems. This work is expected to produce significant breakthrough in nonlinear optimization at large and nonlinear energy optimization in particular.
Broader Impacts: The results of this project can be utilized for the design, modernization, and operation of many power grids around the world. This project will have a significant impact on the power industry by revolutionizing their energymanagement systems, leading to the following immediate benefits: (i) reducing the electricity cost through a cheaper way of dispatching and delivering power, (ii) decreasing the likelihood of power outages by optimizing the reliability and robustness of the grid, and (iii) reducing gas emissions by optimally utilizing the sustainable energy. This project has major outreach activities for a local school, where more than 85% of the student population are significantly underrepresented in STEM-related fields. Management of nonlinearities in real time is especially important to the economics of trying to use more renewable energy sources; thus work on this challenge may be crucial to our ability to accomplish this in an affordable way.
