PDELab: An Object-Oriented Framework for Building Problem Solving Environments for PDE Based Applications

In this paper we present an object-oriented methodology and tools for creating high level, high performance problem solving systems (workbenches) for scientific applications modeled by partial differential equations. This methodology is validated by the creation of a scientific computing workbench for bioseparation analysis. One of the design objectives of PDELab is to provide workbench developers and users with much the same kind of independence in software as they have come to expect in hardware. The adopted architecture of this software platform for creating problem solving environments for PDE applications is devoted to “clean layering.” At the bottom are the various “smart” libraries that support the numerical simulation of various “physical” objects together with the corresponding knowledge bases needed to support the computational intelligence aspects of the various workbenches; at the top is a set of interactive tools that allow the user to carry out his objectives using “natural” tools. Between these layers sits a piece of “middleware” called a “software bus.” Its design objective is to allow the integration of a variety of software components needed to support “hybrid” (numeric and experimental) PDE based workbenches. Moreover, it comes with a software tool that allow its reconfiguration for specific applications. This paper discusses the design and implementation issues of this three layered architecture of PDELab.