On Numerical Error Propagation with Sensitivity
نویسندگان
چکیده
An emerging area of research is to automatically compute reasonably accurate upper bounds on numerical errors, including roundoffs due to the use of a finite-precision representation for real numbers such as floating point or fixed-point arithmetic. Previous approaches for this task are limited in their accuracy and scalability, especially in the presence of nonlinear arithmetic. Our main idea is to decouple the computation of newly introduced roundoff errors from the amplification of existing errors. To characterize the amplification of existing errors, we use the derivatives of functions corresponding to program fragments. We implemented this technique in an analysis for programs containing nonlinear computation, conditionals, and a certain class of loops. We evaluate our system on a number of benchmarks from embedded systems and scientific computation, showing substantial improvements in accuracy and scalability over the state of the art.
منابع مشابه
Estimation of pull-in instability voltage of Euler-Bernoulli micro beam by back propagation artificial neural network
The static pull-in instability of beam-type micro-electromechanical systems is theoretically investigated. Two engineering cases including cantilever and double cantilever micro-beam are considered. Considering the mid-plane stretching as the source of the nonlinearity in the beam behavior, a nonlinear size-dependent Euler-Bernoulli beam model is used based on a modified couple stress theory, c...
متن کاملEstimation of pull-in instability voltage of Euler-Bernoulli micro beam by back propagation artificial neural network
The static pull-in instability of beam-type micro-electromechanical systems is theoretically investigated. Two engineering cases including cantilever and double cantilever micro-beam are considered. Considering the mid-plane stretching as the source of the nonlinearity in the beam behavior, a nonlinear size-dependent Euler-Bernoulli beam model is used based on a modified couple stress theory, c...
متن کاملGlobal Sensitivity Analysis of Ordinary Differential Equations Adaptive Density Propagation Using Approximate Approximations
Ordinary differential equations play an important role in the modeling of many real-world processes. To guarantee reliable results, model design and analysis must account for uncertainty and/or variability in the model input. The propagation of uncertainty & variability through the model dynamics and their effect on the output is studied by sensitivity analysis. Global sensitivity analysis is c...
متن کاملNumerical Modeling of Saline Gravity Currents Using EARSM and Buoyant k- Turbulence Closures
Gravity currents are very common in nature and may appear in rivers, lakes, oceans, and the atmosphere. They are produced by the buoyant forces interacting between fluids of different densities and may introduce sediments and pollutants into water bodies. In this study, the hydrodynamics and propagation of gravity currents are investigated using WISE (Width Integrated Stratified Environments), ...
متن کاملGreen's Function Method for Fast On-Line Learning Algorithm of Recurrent Neural Networks
The two well known learning algorithms of recurrent neural networks are the back-propagation (Rumelhart & el al., Werbos) and the forward propagation (Williams and Zipser). The main drawback of back-propagation is its off-line backward path in time for error cumulation. This violates the on-line requirement in many practical applications. Although the forward propagation algorithm can be used i...
متن کاملLong - Time Error Estimation and a Stability Indicator
Long-time error estimates are abstractly given for a large class of initial value problems without using the traditional concept of \numerical stability". Instead of numerical error propagation, we consider exact error propagation by splitting the error of a numerical initial value problem into local error and propagated global error in a way diierent from the traditional one. The advantage is ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- CoRR
دوره abs/1410.0198 شماره
صفحات -
تاریخ انتشار 2014