Black-box calibration for complex-system simulation
Black-box calibration for complex-system simulation
复制标题
复杂系统仿真的黑盒校准
DOI：
10.1098/rsta.2010.0051
复制
发表时间：
2010
期刊：
影响因子：
--
通讯作者：
Alexander I. J. Forrester
中科院分区：
文献类型：
--
作者：
Alexander I. J. Forrester
预测或测量复杂系统的输出是许多科学领域的重要和具有挑战性的部分。如果参数研究和优化需要多个观测值，则精确的计算密集型预测或昂贵的实验是难以处理的。本文着眼于使用高斯过程为基础的相关性，以纠正简单的计算机模型与稀疏的数据，从物理实验或更复杂的计算机模型。从本质上讲，基于物理的计算机代码和实验被基于快速问题的算法代码所取代。给出了两个气动设计实例。首先，一个廉价的二维势流求解器进行校准，以代表无人机机翼的流动。然后，使用校准后的尾翼模拟来优化赛车的尾翼，以包括整个汽车上的气流影响。
Predicting or measuring the output of complex systems is an important and challenging part of many areas of science. If multiple observations are required for parameter studies and optimization, accurate, computationally intensive predictions or expensive experiments are intractable. This paper looks at the use of Gaussian-process-based correlations to correct simple computer models with sparse data from physical experiments or more complex computer models. In essence, physics-based computer codes and experiments are replaced by fast problem-specific statistics-based codes. Two aerodynamic design examples are presented. First, a cheap two-dimensional potential-flow solver is calibrated to represent the flow over the wing of an unmanned air vehicle. The rear wing of a racing car is then optimized using rear-wing simulations calibrated to include the effects of the flow over the whole car.