Environmental Modeling Research at the University of North Carolina at Chapel Hill: Final Report Page: 3 of 6
This report is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
The specific features and emphases of PSE's vary with the application domain and
target audience. However, there are two common approaches that can be used to
characterize PSE's. The first is a library-based approach. As its name suggests, a
library-based approach provides a simulation environment built on a substantial
kernel of existing library routines providing a range of solution methods and
functionality. The user then orchestrates the library routines inside the PSE's
simulation environment to solve his or her problem.
This approach has the drawback that it does not scale very well: as the range of
problems and sophistication of solution methods increases, the interaction
between methods becomes more complex and simple compositionality of the
library routines is easily lost. Moreover an increasingly large burden is placed on
the user of such a system to formulate even simple problems.
On the other hand, a software synthesis, or transformational, approach constructs
a simulator directly from the high-level problem description given by the user.
Producing a simulator using this approach involves the application of a set of
transformations by which algorithms are introduced and data structures are
refined from an abstract, high level into a concrete implementation to be executed
on a given platform.
To provide a simple example, consider a problem posed as a set of ordinary
differential equations (ODE's). Depending on the form of the equations, the most
appropriate solver may be an implicit ODE solver, an explicit ODE solver, or a
differential algebraic equation (DAE) solver. The library approach would provide
each of these solvers, but the PSE user would have to construct the appropriate
function arguments. The arguments required are quite different for the three
solvers, and must be extracted from the problem description. A synthesis
approach automates this process, and may automate (or at least simplify) the
selection of the appropriate method, by inspection of the form of the ODE. The
synthesis approach is complex, and is only feasible for a limited domain of
problems. Thus its applicability will be more restricted than a library-based
Our initial work focused on evaluating the relative merits of a library or
transformational approach for developing a PSE. We decided to take a hybrid
approach that offers the ability to refine a problem specification from an initial
high-level description but is not required to generate the entire set of solution
methods as would typically be done in a transformational approach. The key to
realizing this goal is to provide sufficiently rich internal representations of the
underlying mathematical problem as well as the algorithms and solution methods
necessary to achieve a solution. Using an initial prototype, solutions to the first
Here’s what’s next.
This report can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Report.
Miller, C.T. Environmental Modeling Research at the University of North Carolina at Chapel Hill: Final Report, report, December 19, 2006; United States. (digital.library.unt.edu/ark:/67531/metadc884968/m1/3/: accessed December 17, 2017), University of North Texas Libraries, Digital Library, digital.library.unt.edu; crediting UNT Libraries Government Documents Department.