Uncertainty Management in Computer Assisted Problem Solving Environment (PSE) for Scientific Computing

Abstract

Abstract?Uncertainty in scientific computing is discussed in this paper. Uncertainty comes from various sources, for example, from physical model ncertainty, mathematical model errors, unknown input data, numerical model errors, insufficient numerical precision, round-off error, floating point precision, rogramming errors, data processing errors or uncertainty, visualization errors, etc., as well as human errors concerned. The uncertainty in scientific computing has induced serious accidents or disasters, so far. The uncertainty in scientific computing must be managed to avoid serious accidents anddisasters, and also to obtain reasonable or precise scientific results for discoveries and scientific progresses. The uncertainty management is a hard job in scientific computing. On the other hand, studies on problem solving environments (PSEs) have been explored to enhance programming power and to help users to solve problems using computers. PSE is defined as “a system that provides all the computational facilities necessary to solve a target class of problems. It uses the language of the target class and users need not have specialized knowledge of the underlying hardware or software”. Theuncertainty management could be supported by the concept of PSE. In this paper first we discuss on the uncertainty properties and their origin in scientific computing. Then discussed is how to support the uncertainty management by the PSE concept.