• Performs non-linear optimization, parameter estimation, and regression
• Arbitrary functional forms can be optimized or regressed against data
• High dimension parameters, experiments, and independent variables accommodated
• Accommodates arbitrary common or missing data structure
• Automatically builds appropriately formatted workspaces for data IO and equations
• Flexible optimal combination of methods for optimization, line-search, and search accuracy
• Variety of optimization strategies; including full Hessian second derivatives, as well as lesser order methods including steepest descent, two variations of conjugate gradient, and Gauss-Newton in the case of regression
• Fast and original line-search up to 5th order, with or without accuracy control
• Fast and robust and extensively benchmarked; avoids bogging down or losing control in high parameter correlation, or for convex-negative and or kinked surfaces
• Employs Choleski decomposition as appropriate to ensure efficiency
• Computes confidence intervals of parameters, with correlation matrix
• Friendly and intuitive user interface
• Graphical depiction of 2-parameter space paths during optimization

736 Hollow Rd, Ellicott City, MD 21043
443-546-7157
Questions? kfroenigk@gmail.com
Copyright © 2021 Karl Roenigk All rights reserved.