PEST - Model-Independent Parameter Estimation and Uncertainty Analysis

The Next Generation

PEST++ and GENIE

PEST++ is a model-independent parameter estimator. It uses the same means as PEST does for communicating with a model - i.e. template and instruction files. As presently programmed it also reads a PEST control file; hence its use is interchangeable with that of PEST.

PEST++ was written by Dave Welter. As the name implies it is written in C++. It is object oriented and modular; hence its capabilities can be readily expanded over time.

The initial version of PEST++ is powerful. It was built with highly-parameterized inversion in mind. It implements Tikhonov regularization and subspace methods to achieve well-posedness of nonunique inverse problems. It also implements the SVD-Assist methodology of PEST, but in a more flexible way.

GENIE is a generalized parallel run manager. It was written by Chris Muffels and Willem Shreuder (Willem is the force behind BEOPEST). It uses TCP/IP for communication between master and slaves, but does this in a more flexible and robust way than BEOPEST.

GENIE is available as a module that can be called by any program. Hence its use is not limited to PEST++. In fact any application that needs to run a model many times, and (optionally) uses template and instruction files to communicate with that model, can use GENIE. Any FORTRAN or C programmer can use GENIE to parallelize model runs across cores in a single machine, machines in an office network, or networks in different parts of the world. 

To find out more about PEST++ and GENIE, and to download them, go to this site .

C4SF Login | Web Site Managed via C4Site Factory,™ a trademark of Echo Valley Graphics, Inc.
©2014 Copyright Echo Valley Graphics, Inc. All Rights Reserved.