James Leek | 18-ERD-025
Parallel Discrete Events Simulation (PDES) is used to model complex asynchronous systems, such as computer networks, satellite systems, vehicular traffic, and human performance. Although uncertainty quantification (UQ) techniques have become a mainstay of verification and validation efforts on physics simulations, PDES models are very different from traditional physics simulations, and research into UQ techniques for PDES is in its infancy. It is not clear which traditional UQ techniques can be applied to PDES, or what new techniques will need to be developed.
The goal of this project was to identify existing UQ techniques that can be applied to PDES and to develop new techniques as necessary. We implemented or developed techniques to handle issues that do not appear in traditional physics simulations but are common among PDES, including sampling techniques for high-dimensional homogenous inputs, response surfaces for high-variance heteroskedastic output, and characterization of skewed output distributions. We developed new, efficient sampling and response surface methods, and we demonstrated that the output can be characterized efficiently and that basic UQ can be performed. These foundational UQ techniques are needed to conduct UQ analyses, including novel applications of UQ at Lawrence Livermore National Laboratory.
Our results lay the groundwork for novel applications of uncertainty quantification to PDES projects that are an important part of the Laboratory's national security mission, including modeling complex satellite systems. In addition, this work supports the Laboratory's core competency in high-performance computing, simulation, and data science.
Publications, Presentations, and Patents
Leek, J. 2018. "Application of Uncertainty Quantification to Discrete Event Simulations." Lawrence Livermore National Laboratory, Computation Engineering Division (CED) Seminar, November 2018. LLNL-PRES-761682
Quinlan, K. R. 2019a. "Uncertainty Quantification for Parallel Discrete Event Simulation." Statistical Perspectives on Uncertainty Quantification Conference, Raleigh, NC, May 2019. LLNL-POST-773060
——— 2019b. "Uncertainty Quantification for Parallel Discrete Event Simulation. "Joint Statistical Meetings. Denver, CO, July 2019. LLNL-PRES-781760
——— 2020b. "Uncertainty Quantification for Parallel Discrete Event Simulation." Conference on Data Analysis, Santa Fe, NM, February 2020. LLNL-POST-805480
Quinlan, Kevin R., and Jim R. Leek. 2020a. "quantkriging: Quantile Kriging for Stochastic Simulations with Replication." CRAN. LLNL-CODE-796243
———2020b. "simplexdesign: Simplex Design for Stochastic Simulations and Agent Based Models." LLNL-CODE-796317
Tong, C. 2018. "Uncertainty Quantification for Discrete Event Simulations." Lawrence Livermore National Laboratory, CED Seminar, November 2018. LLNL-PRES-761621