|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||Computational Design and Simulation of Materials (CDSM 2018): Computational Design of Materials
||On the Fly Efficient Global Optimization Techniques to Accelerate Materials Design
||Anjana Talapatra, Shahin Boluki, Thien Duong, Raymundo Arroyave, Xiaoning Qian, Edward Dougherty
|On-Site Speaker (Planned)
Advances in accelerating materials discovery have been characterized by a data glut; which do not always translate into knowledge. As shown by Lookman et al., a targeted approach using adaptive design strategies is a more 'intelligent' way to approach the problem of materials design. Inspired by this, an on-the-fly, real-time implementation of the Effective Global Optimization (EGO) algorithm to answer real-world materials questions is presented. We solve some single and multi-objective design problems to showcase the efficiency of this approach. We integrate this framework with Bayesian Model Averaging to incorporate model uncertainty wherein the weight of different models is calculated based on the prior probability and likelihood of the observed data thereby leading to autonomous model selection. In spite of scant initial data and the absence of good models, it is possible to leverage these methods to reach a design solution optimally via a minimum number of experiments and/or simulations.
||Planned: Supplemental Proceedings volume