An Efficient and Accurate Building Optimization Strategy Using Singular Value Decomposition

摘要

Optimizing the life cycle cost of a building typically involves a large number of variables due to the many options that exist at the time that a building is being designed. Such large-scale optimization problems are often prohibitive within the building industry because of the excessive computational time required by the building energy modeling software; therefore, any optimization studies that are performed during a building design are typically only completed using a small number of variables. To achieve the goal of performing a life cycle building optimization in an acceptable time frame, this paper proposes an accurate and efficient method using singular value decomposition on the design variables. Through the use of singular value decomposition a large number of design variables can be reduced to a smaller subset of design variables that can be solved more quickly by the optimization algorithm. In this paper the authors apply this methodology to a case study of a typical residential building in six separate locations across the U.S. and compare the results with those of the full optimization process over the entire design space.