LSTSQ: a module for reliable constrained and unconstrained nonlinear regression.

A software module for nonlinear regression analysis, based on the reliable Meyer-Roth algorithm (a modified damped least square algorithm), is presented. It allows both constrained and unconstrained optimization, and the use of a variety of weighting methods. Virtually any nonlinear function can be fitted, including those with several nondependent variables. The package has been thoroughly tested, and is available in several common computer languages (Pascal, Modula-2 and C). It is easy to use, and advanced knowledge of mathematics or computers is not essential. Standard test problems, and a fully working example of use on enzyme kinetics are included.