|
Nonlinear
Regression
Example
Problems
Screen
Shots
Download
Order
|
|
Auto2Fit is a robust, flexible,
easiest-to-use global optimization, nonlinear regression, equation solve and parameter estimation
program with the start-of-the-art optimization a;gorithms. It is designed especially for general users to deal with
the complicated jobs that are hard or impossible to be solved by
regular/traditional approaches. Its intuitive, simple
and friendly user-interface, straightforward operations
and high efficient performance make it the most powerful tool for science and
engineering study.
Return
to top
2: What can be done with Auto2Fit?
Auto2Fit can be used in lots
of study areas such as:
- Regression and Curve Fit: linear,
nonlinear, multi-nonlinear with no limitations on variable and parameter
number. With its revolutional Gobal Levenberg-Marquardt (GLM),
for example, all the Statistical Reference Datasets Project (SRDP) of the National Institute of
Standards and Technology (NIST)
can be solved successfully without using
initial start values provided by NIST, but free
random values.
- Global Optimization: find
the maximal/minimal value of any function. The function may be nonlinear,
non continues, non derivable.
- Equation Solve: solve any
type of equation or system equations
- Model
Auto-calibration: parameter estimation of complicated engineering model
Return to top
Eight types of optimization
algorithms as well as many of their
improved/modified versions and new schemes have been implemented Auto2Fit. It
therefore provides plenty of choices for end users to meet various
problems.
- Levenberg-Marquardt (LM):
standard LM and Global LM
- Quasi-Newton (BFGS):
standard BFGS and Global BFGS
- Simplex Method (SM)
- Differential Evolution (DE): ten schemes (5
new) and four routs (3 new)
- Genetic Algorithm (GA): both encode/decode
mode and real-value based mode, six crossover types and seven selection types
- Particle Swarm Optimization (PSO): six
schemes (2 standard and 4 new)
- Maximum Inherit Optimization (MIO): four schemes
- Simulated Annealing (SA)
- Self-Organizing
Migrating Algorithms (SOMA)
- Conjugate-Gradient
Method (CGM)
- Powell Optimization
(OP)
- Tabu Search (TS)
- Simplex Linear Programming
Auto2Fit solve all problems in two ways:
Quick Mode and Program Mode
-
Quick Mode: simply input one or more string expressions to represent/describe the
problems. This is suitable for most of cases.
- Program
Mode: in the case that the problems are impossible to be described by a simple
string expression, then you may use Pascal/Delphi or VB style script to
describe the problem
More features:
- The script engine used in Auto2Fit is extremely
fast, for example, 5 - 30 times faster than VB
- Multi-data file, multi-outputs
supported: this unique feature makes it very useful for parameter estimation
with multi data file, for example, hydrological modelling.
- Easy to handle file-depended
parameters and file-depended constants
- Excel-like, multi-windows spreadsheet
embedded for easy pre- and post-data process
- File tree shell provides great convenient
for accessing data file
- Lots of examples, including some
classical optimization problems
Return to top
Running the 32-bit
version of Auto2Fit requires a minimum 486 processor and Windows 95/98/2000/ME/XP/Vista. A complete installation requires less than 10 MB of hard disk space.
We welcome your input on how we can improve NeuralPower to make it more useful
to the users.
Return to top
)
|