We investigate the behavior of adaptive time stepping numerical algorithms under the reverse mode of automatic differentiation (AD). By differentiating the time step controller and the error estimator of the original algorithm, reverse mode AD generates spurious adjoint derivatives of the time steps. The resulting discrete adjoint models become inconsistent with the adjoint ODE, and yield incor...

A framework for real-time communication over a wireless channel is proposed. The concept of deadline dependent coding (DDC), previously suggested by the authors, is further developed using soft decision decoding of block codes to maximize the probability of delivering information before a given deadline. The strategy of DDC is to combine different coding and decoding methods with automatic repe...

This paper discusses a formal, mathematically-based, approach to the analysis of operator interaction with machines, in general, and with complex and automated control systems, in particular. It addresses the problem of correctness of displays by asking whether the display provides the necessary information about the machine to enable the operator to perform a specified task successfully and un...

We have developed a menu-selection-based Japanese input method for a pen device. To obtain user-input kanji-form text rapidly, the system requires sequence of consonant. Following this, the user selects kanji-form candidates inferred by the system. In some situations, consonant sequences can trigger an explosion of kanji-form candidates. Therefore, we have implemented a method for reducing cand...

3 Example: Differentiating the Lotka-Volterra ODE system 2 3.1 The Lotka-Volterra ODE system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 3.2 The derivatives of the Lotka-Volterra ODE system . . . . . . . . . . . . . . . . . . 3 3.3 Integration of the original and augmented Lotka-Volterra ODE system with the Runge-Kutta method . . . . . . . . . . . . . . . . . . . . . . . . . . . ....

Sensitivity analysis is a research area of increasing interest within CFD. Recent developments in Automatic Differentiation software (AD) make the discrete adjoint methods particularly attractive. We outline the theory of adjoint sensitivities and discuss how to use AD to derive adjoint versions from CFD codes. In particular we review the properties of the four major AD packages, we explain how...

This paper explains precisely how a declarative method language, based upon the formal notations of Z and B, can be used as a basis for automatic code generation. The language is used to describe the intended effect of operations, or methods, upon the components of an object model; each method is defined by a pair of predicates—preand post-conditions. Following the automatic incorporation of mo...