Multi-system Gaussian Process based Surrogate Modeling and Sequential Design for Expensive Computer Simulations  
Author Jiayu Dong

 
Co-Author(s) Jianguo Wu

 
Abstract Computer simulations have emerged as potent tools for understanding physical laws and optimizing design solutions. It facilitates computer programs to virtually approximate the running of production systems and the mechanism of some physical phenomena without the need for expensive lab experiments, which significantly economizes on experimental resources and time. To augment the reliability and precision of simulation outputs, however, the complexity of simulation systems has markedly increased, accompanied by a sharp surge in computational resource demand. Unfortunately, such a huge demand cannot be entirely meet by current computer technology levels. To address this challenge, surrogate model was proposed and extensively investigated.

 
Keywords Multiple simulation systems, Surrogate model, Gaussian process, Statistical learning, Sequential design
   
    Article #:  DSBFI25-21
 
Proceedings of 3rd ISSAT International Conference on Data Science in Business, Finance and Industry
January 6-8, 2025 - Da Nang, Vietnam