TY - GEN
T1 - A novel machine-learning aided optimization technique for material design
T2 - ASME 2016 Heat Transfer Summer Conference, HT 2016, collocated with the ASME 2016 Fluids Engineering Division Summer Meeting and the ASME 2016 14th International Conference on Nanochannels, Microchannels, and Minichannels
AU - Hajimirza, Shima
PY - 2016
Y1 - 2016
N2 - Patterned thin film structures can offer spectrally selective radiative properties that benefit many engineering applications including photovoltaic energy conversion at extremely efficient scales. Inverse design of such structures can be expressed as an interesting optimization problem with a specific regime of complexity; namely moderate number of optimization parameters but highly time-consuming forward problem. For problems like this, a search technique that can somehow learn and parameterize the multi-dimensional behavior of the objective function based on past search points can be extremely useful in guiding the global search algorithm and expediting the solution for such complexity regimes. Based on this idea, we have developed a novel search algorithm for optimizing absorption coefficient of visible light in a multi-layered siliconbased nano-scale thin film solar cell. The proposed optimization algorithm uses a machine-learning predictive tool called regression-tree in an intermediary step to learn (i.e. regress) the objective function based on a previous generation of random search points. The fitted model is then used as a guide to resample from a new generation of candidate solutions with a significantly higher average gain. This process can be repeated multiple times and better solutions are obtained with high likelihood at each stage. Through numerical experiments we demonstrate how in only one resampling stage, the propose technique dominates the state-of-the-art global search algorithms such as gradient based techniques or MCMC methods in the considered nano-design problem.
AB - Patterned thin film structures can offer spectrally selective radiative properties that benefit many engineering applications including photovoltaic energy conversion at extremely efficient scales. Inverse design of such structures can be expressed as an interesting optimization problem with a specific regime of complexity; namely moderate number of optimization parameters but highly time-consuming forward problem. For problems like this, a search technique that can somehow learn and parameterize the multi-dimensional behavior of the objective function based on past search points can be extremely useful in guiding the global search algorithm and expediting the solution for such complexity regimes. Based on this idea, we have developed a novel search algorithm for optimizing absorption coefficient of visible light in a multi-layered siliconbased nano-scale thin film solar cell. The proposed optimization algorithm uses a machine-learning predictive tool called regression-tree in an intermediary step to learn (i.e. regress) the objective function based on a previous generation of random search points. The fitted model is then used as a guide to resample from a new generation of candidate solutions with a significantly higher average gain. This process can be repeated multiple times and better solutions are obtained with high likelihood at each stage. Through numerical experiments we demonstrate how in only one resampling stage, the propose technique dominates the state-of-the-art global search algorithms such as gradient based techniques or MCMC methods in the considered nano-design problem.
UR - http://www.scopus.com/inward/record.url?scp=85002905653&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85002905653&partnerID=8YFLogxK
U2 - 10.1115/HT2016-7306
DO - 10.1115/HT2016-7306
M3 - Conference contribution
AN - SCOPUS:85002905653
T3 - ASME 2016 Heat Transfer Summer Conference, HT 2016, collocated with the ASME 2016 Fluids Engineering Division Summer Meeting and the ASME 2016 14th International Conference on Nanochannels, Microchannels, and Minichannels
BT - Heat Transfer in Multiphase Systems; Gas Turbine Heat Transfer; Manufacturing and Materials Processing; Heat Transfer in Electronic Equipment; Heat and Mass Transfer in Biotechnology; Heat Transfer Under Extreme Conditions; Computational Heat Transfer; Heat Transfer Visualization Gallery; General Papers on Heat Transfer; Multiphase Flow and Heat Transfer; Transport Phenomena in Manufacturing and Materials Processing
Y2 - 10 July 2016 through 14 July 2016
ER -