DDAPLUS:
Differential-Algebraic
Solver and sensitivity analyzer for nonlinear initial-value problems
in one dimension; also solves algebraic nonlinear equation systems.
DDAPLUS is an extension of the solver DASSL (Petzold, 1982),
considerably enhanced since the version of Caracotsios and Stewart,
reference [2]. DDAPLUS provides superior initialization and no
negativity control, stricter error control and reporting, optional
optimization of the initial step size (even when the initial
derivatives are zero), and the ability to stop and report at a
target value of any designated variable. The basic algorithms are
summarized in references [1], [2] and [3], listed on the following
page.
GREGPLUS:
A major enhancement of
our nonlinear parameter estimation code GREG for process model
analysis and sequential experimental design. Estimates the model
parameters, their inference intervals and covariance, using
single-response or multi-response data. Gives similar estimates for
auxiliary functions of the parameters. Compares rival models, using
replicate data along with objective criteria of posterior
probability and goodness of fit. The code includes weighted
nonlinear least squares, the multiresponse objective functions of
Box and Draper [4] and Bard [5], and further developments from
references [6], [7] and [8]. Diagnostics and corrective actions for
over parameterized models are provided, including a display of the
test divisors that determine which parameters are estimable
according to the user's pivoting tolerance. Uninformative response
variables are identified and set aside. Parametric sensitivities for
the estimation can be coded by the user or can be generated
accurately by GREG Plus’s new adaptive finite-difference algorithms.
Next
===>