CONFIRMED: EAS Doctoral Proposal Defense by Esmaeil Rezaei
EAS Doctoral Proposal Defense by Esmaeil Rezaei Date: Friday, December 15, 2023 Time: 10:00 a.m. Topic: Towards better electricity supply management: a reduced-basis enabled dimension reduction for demand forecasting using smart-meter big data Location: CSCDR - TXT 105 Abstract: Real-world datasets often comprise a large number of features or variables, encompassing billions of samples. Manipulating such high-dimensional data is computationally expensive. Although some advanced strategies leverage high performance and parallel computing capabilities, this could entail complex and costly resources for users. In response, dimension reduction techniques have been developed that transform high-dimensional data into a low-dimensional space while preserving properties critical to the original data. Dimension reduction has found widespread use across various domains, including pattern recognition, clustering, classification, and accelerating numerical simulations of complex physical phenomena. Regardless of the strategy adopted, available techniques heavily rely on computationally intensive matrix factorizations, such as Singular Value Decomposition (SVD). These techniques are thus quickly becoming intractable as the world adapts to a new norm where massive datasets have become everyday commodities. In this research, we introduce a novel dimension reduction strategy, HYBRID: HYper-reduced Basis Reduction via Interactive Decomposition, which offers remarkable efficiency and includes an error indicator certifying the accuracy by which the properties of the original data are preserved. HYBRID draws upon reduced basis decomposition and recent advancements in Reduced Basis Methods (RBMs), which have garnered attention as efficient dimension reduction tools for solving parametrized partial differential equations. Specifically, we propose speeding up the construction of the reduced basis by adopting the concept of Reduced Residual, enabling efficient error measurement on a subset of dimensions proportional to the intrinsic dimension of the given dataset. We present numerical examples to demonstrate the performance of the proposed HYBRID technique and its competitive edges over exiting dimension reduction techniques. Additionally, we apply HYBRID to a large-scale utility dataset encompassing 3.7 million customers of Commonwealth Edison (ComEd) company in the state of Illinois to accurately forecast demand. Such accurate forecasting is vital for developing resilient management strategies for energy systems. In a highly competitive era, securing the largest market share through high-quality services demands meticulous planning, particularly in accurately forecasting future loads to prevent shortages. While oversupplying with fuel can satisfy demand, the costs of energy storage technology are prohibitively high. Therefore, effective and purposeful planning is necessary for companies to ensure precise load forecasting, crucial in maintaining the reliability and resilience of electricity services. ADVISOR(S): Dr. Mazdak Tootkaboni, Dept of Civil and Environmental Engineering (Advisor) (mtootkaboni@umassd.edu) Dr. Arghavan Louhghalam, Dept of Civil and Environmental Engineering (Co-Advisor) (Arghavan_Louhghalam@uml.edu) COMMITTEE MEMBERS: Dr. Yanlai Chen, Department of Mathematics Dr. Alfa Heryudono, Department of Mathematics Dr. Negin Alemazkoor, Dept of Civil and Environmental Engineering, UVA NOTE: All EAS Students are ENCOURAGED to attend.
See description for location
Engineering and Applied Sciences
easphd@umassd.edu