Reliable computing

Showing results in 'Publications'. Show all posts
Dawood, Hend. InCLosure (Interval enCLosure): A Language and Environment for Reliable Scientific Computing. 3.0 ed. Department of Mathematics, Faculty of Science, Cairo University, 2020. AbstractWebsite

InCLosure (Interval enCLosure) is a Language and Environment for Reliable Scientific Computing. Interval computations are radically different from traditional numerical approximation methods and reliable computing under uncertainty is a key focus for modern research in mathematics, computer science, physics, and engineering. InCLosure is a system for carrying out reliable and self-validated computations in arbitrary precision. From its name, InCLosure (abbreviated as "InCL") focuses on "computing guaranteed interval enclosures", that is, "enclosing the exact real result in an interval".
InCLosure is powerful enough to carry out computations ranging from simple real and interval arithmetic, through symbolic and numeric differentiation to an arbitrary order, real automatic differentiation, and interval automatic differentiation, up to and including interval enclosures of integrals and Taylor model computations. No matter how complicated the problem under consideration is, InCLosure provides arbitrary precisions and reliable interval results that can be as narrow as possible by the computational power of the hosting machine.
InCLosure supports arbitrary precision in both real and interval computations. In real arithmetic, the precision is arbitrary in the sense that it is governed only by the computational power of the machine (default is 20 significant digits). The user can change the default precision according to the requirements of the application under consideration. Since interval arithmetic is defined in terms of real arithmetic, interval computations inherit the arbitrary precision of real arithmetic with an added property that the interval subdivision method is provided with an arbitrary number of subdivisions which is also governed only by the computational power of the machine. The user can get tighter and tighter guaranteed interval enclosures by setting the desired number of subdivisions to cope with the problem at hand.
All the computations defined in terms of real and interval arithmetic (e.g., real and interval automatic differentiation, Taylor model computations, and so forth) inherit the same arbitrary precision.
InCLosure is designed to support both interactive and batch modes. In the InCLosure interactive interface, the user can input an InCL command and see its result before moving on to the next command. InCLosure can also be used in batch mode in which case sequences of InCL commands can be given to InCLosure via InCL input files with the results saved in simple and intuitively formatted output files.
InCLosure is coded entirely in Lisp, arguably the fastest and most powerful language for scientific computations. InCLosure provides a friendly and easy-to-use user interface, a simple and intuitive language, a detailed documentation, and clear and fast results. Anyone can compute with InCLosure.

Dawood, Hend, and Yasser Dawood. "Parametric Intervals: More Reliable or Foundationally Problematic?" Online Mathematics Journal 1, no. 3 (2019): 37-54. Abstractomj_01-03_p37-54_dawood.pdfWebsite

Interval arithmetic has been proved to be very subtle, reliable, and most fundamental in addressing uncertainty and imprecision. However, the theory of classical interval arithmetic and all its alternates suffer from algebraic anomalies, and all have difficulties with interval dependency. A theory of interval arithmetic that seems promising is the theory of parametric intervals. The theory of parametric intervals is presented in the literature with the zealous claim that it provides a radical solution to the long-standing dependency problem in the classical interval theory, along with the claim that parametric interval arithmetic, unlike Moore's classical interval arithmetic, has additive and multiplicative inverse elements, and satisfies the distributive law. So, does the theory of parametric intervals accomplish these very desirable objectives? Here it is argued that it does not.

Keywords: Interval mathematics, Classical interval arithmetic, Parametric interval arithmetic, Constrained interval arithmetic, Overestimation-free interval arithmetic, Interval dependency, Functional dependence, Dependency predicate, Interval enclosures, S-semiring, Uncertainty, Reliability.

Dawood, Hend. "On Some Algebraic and Order-Theoretic Aspects of Machine Interval Arithmetic." Online Mathematics Journal 1, no. 2 (2019): 1-13. Abstractomj_01-02_p1-13_dawood.pdfWebsite

Interval arithmetic is a fundamental and reliable mathematical machinery for scientific computing and for addressing uncertainty in general. In order to apply interval mathematics to real life uncertainty problems, one needs a computerized (machine) version thereof, and so, this article is devoted to some mathematical notions concerning the algebraic system of machine interval arithmetic. After formalizing some purely mathematical ingredients of particular importance for the purpose at hand, we give formal characterizations of the algebras of real intervals and machine intervals along with describing the need for interval computations to cope with uncertainty problems. Thereupon, we prove some algebraic and order-theoretic results concerning the structure of machine intervals.

keywords: Interval mathematics, Machine interval arithmetic, Outward rounding, Floating-point arithmetic, Machine monotonicity, Dense orders, Orderability of intervals, Symmetricity, Singletonicity, Subdistributive semiring, S-semiring.

Dawood, Hend, and Yasser Dawood. "A Logical Formalization of the Notion of Interval Dependency: Towards Reliable Intervalizations of Quantifiable Uncertainties." Online Mathematics Journal 1, no. 3 (2019): 15-36. Abstractomj_01-03_p15-36_dawood.pdfWebsite

Progress in scientific knowledge discloses an increasingly paramount use of quantifiable properties in the description of states and processes of the real-world physical systems. Through our encounters with the physical world, it reveals itself to us as systems of uncertain quantifiable properties. One approach proved to be most fundamental and reliable in coping with quantifiable uncertainties is interval mathematics. A main drawback of interval mathematics, though, is the persisting problem known as the "interval dependency problem". This, naturally, confronts us with the question: Formally, what is interval dependency? Is it a meta-concept or an object-ingredient of interval and fuzzy computations? In other words, what is the fundamental defining properties that characterize the notion of interval dependency as a formal mathematical object? Since the early works on interval mathematics by John Charles Burkill and Rosalind Cecily Young in the dawning of the twentieth century, this question has never been touched upon and remained a question still today unanswered. Although the notion of interval dependency is widely used in the interval and fuzzy literature, it is only illustrated by example, without explicit formalization, and no attempt has been made to put on a systematic basis its meaning, that is, to indicate formally the criteria by which it is to be characterized. Here, we attempt to answer this long-standing question. This article, therefore, is devoted to presenting a complete systematic formalization of the notion of interval dependency, by means of the notions of Skolemization and quantification dependence. A novelty of this formalization is the expression of interval dependency as a logical predicate (or relation) and thereby gaining the advantage of deducing its fundamental properties in a merely logical manner. Moreover, on the strength of the generality of the logical apparatus we adopt, the results of this article are not only about classical intervals, but they are meant to apply also to any possible theory of interval arithmetic. That being so, our concern is to shed new light on some fundamental problems of interval mathematics and to take one small step towards paving the way for developing alternate dependency-aware interval theories and computational methods.

Keywords: Interval mathematics; Interval dependency; Functional dependence; Skolemization; Guaranteed bounds; Interval enclosures; Interval functions; Quantifiable uncertainty; Scientific knowledge; Reliability; Fuzzy mathematics; InCLosure.

Dawood, Hend. InCLosure (Interval enCLosure): A Language and Environment for Reliable Scientific Computing. 1.0 ed. Department of Mathematics, Faculty of Science, Cairo University, 2018. AbstractWebsite

InCLosure (Interval enCLosure) is a Language and Environment for Reliable Scientific Computing. InCLosure, provides rigorous and reliable results in arbitrary precision. From its name, InCLosure (abbreviated as "InCL") focuses on "enclosing the exact real result in an interval". The interval result is reliable and can be as narrow as possible.
InCLosure supports arbitrary precision in both real and interval computations. In real arithmetic, the precision is arbitrary in the sense that it is governed only by the computational power of the machine (default is 20 significant digits). The user can change the default precision according to the requirements of the application under consideration. Since interval arithmetic is defined in terms of real arithmetic, interval computations inherit the arbitrary precision of real arithmetic with an added property that the interval subdivision method is provided with an arbitrary number of subdivisions which is also governed only by the computational power of the machine. The user can get tighter and tighter guaranteed interval enclosures by setting the desired number of subdivisions to cope with the problem at hand.
All the computations defined in terms of real and interval arithmetic (e.g., real and interval automatic differentiation) inherit the same arbitrary precision.
InCLosure is written in Lisp, the most powerful and fast language in scientific computations. InCLosure provides easy user interface, detailed documentation, clear and fast results. Anyone can compute with InCLosure.

Dawood, Hend. "Interval Mathematics as a Potential Weapon against Uncertainty." In Mathematics of Uncertainty Modeling in the Analysis of Engineering and Science Problems. Hershey, PA: IGI Global, 2014. Abstractinterval_mathematics_as_a_potential_weapon_against_uncertainty.pdf

This chapter is devoted to introducing the theories of interval algebra to people who are interested in applying the interval methods to uncertainty analysis in science and engineering. In view of this purpose, we shall introduce the key concepts of the algebraic theories of intervals that form the foundations of the interval techniques as they are now practised, provide a historical and epistemological background of interval mathematics and uncertainty in science and technology, and finally describe some typical applications that clarify the need for interval computations to cope with uncertainty in a wide variety of scientific disciplines.

Keywords: Interval mathematics, Uncertainty, Quantitative Knowledge, Reliability, Complex interval arithmetic, Machine interval arithmetic, Interval automatic differentiation, Computer graphics, Ray tracing, Interval root isolation.

Dawood, Hend. Interval Mathematics: Foundations, Algebraic Structures, and Applications. Cairo: Cairo University, 2012. Abstractinterval_mathematics_msc_thesis_by_hend_dawood.pdf

We begin by constructing the algebra of classical intervals and prove that it is a nondistributive abelian semiring. Next, we formalize the notion of interval dependency, along with discussing the algebras of two alternate theories of intervals: modal intervals, and constraint intervals. With a view to treating some problems of the present interval theories, we present an alternate theory of intervals, namely the "theory of optimizational intervals", and prove that it constitutes a rich S-field algebra, which extends the ordinary field of the reals, then we construct an optimizational complex interval algebra. Furthermore, we define an order on the set of interval numbers, then we present the proofs that it is a total order, compatible with the interval operations, dense, and weakly Archimedean. Finally, we prove that this order extends the usual order on the reals, Moore's partial order, and Kulisch's partial order on interval numbers.

Keywords:
Classical interval arithmetic, Machine interval arithmetic, Interval dependency, Constraint intervals, Modal intervals, Classical complex intervals, Optimizational intervals, Optimizational complex intervals, S-field algebra, Ordering subsets of the reals, Interval arithmetic, Ordering interval numbers, Ordinal power, Total order, Well order, Order compatibilty, Weak Archimedeanity, Dedekind completeness, Interval lattice, Interval order topology, Moore's partial order, Kulisch's partial order.

Dawood, Hend. Theories of Interval Arithmetic: Mathematical Foundations and Applications. Saarbrücken: LAP Lambert Academic Publishing, 2011. Abstractexcerptoftheoriesofintervalarithmeticlap2011-henddawood.pdfWebsite

Reviews

"This new book by Hend Dawood is a fresh introduction to some of the basics of interval computation. It stops short of discussing the more complicated subdivision methods for converging to ranges of values, however it provides a bit of perspective about complex interval arithmetic, constraint intervals, and modal intervals, and it does go into the design of hardware operations for interval arithmetic, which is something still to be done by computer manufacturers."
   - Ramon E. Moore, (The Founder of Interval Computations)
   Professor Emeritus of Computer and Information Science, Department of Mathematics, The Ohio State University, Columbus, U.S.A.

"A popular math-oriented introduction to interval computations and its applications. This short book contains an explanation of the need for interval computations, a brief history of interval computations, and main interval computation techniques. It also provides an impressive list of main practical applications of interval techniques."
   - Vladik Kreinovich, (International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems)
   Professor of Computer Science, University of Texas at El Paso, El Paso, Texas, U.S.A.

"I am delighted to see one more Egyptian citizen re-entering the field of interval mathematics invented in this very country thousands years ago."
   - Marek W. Gutowski,
   Institute of Physics, Polish Academy of Sciences, Warszawa, Poland

Book Description

Scientists are, all the time, in a struggle with uncertainty which is always a threat to a trustworthy scientific knowledge. A very simple and natural idea, to defeat uncertainty, is that of enclosing uncertain measured values in real closed intervals. On the basis of this idea, interval arithmetic is constructed. The idea of calculating with intervals is not completely new in mathematics: the concept has been known since Archimedes, who used guaranteed lower and upper bounds to compute his constant Pi. Interval arithmetic is now a broad field in which rigorous mathematics is associated with scientific computing. This connection makes it possible to solve uncertainty problems that cannot be efficiently solved by floating-point arithmetic. Today, application areas of interval methods include electrical engineering, control theory, remote sensing, experimental and computational physics, chaotic systems, celestial mechanics, signal processing, computer graphics, robotics, and computer-assisted proofs. The purpose of this book is to be a concise but informative introduction to the theories of interval arithmetic as well as to some of their computational and scientific applications.