Nelder-Mead Simplex Solver with Robust Input Options

General form implementation of a downhill Amoeba optimization algorithm, accepting a function input which describes fit.

General form implementation of a downhill Amoeba optimization algorithm accepting a function input which describes fit, freeing the user from having to build the matrices. This method is best at finding local minima, and boundary conditions or initial conditions can change the result converged to. If the performance function is not continuous and smooth, or has multiple local minima, the simplex method may not produce desirable results.

Simple boundary conditions can be handled with the lowBoundary and highBoundary optional inputs, which can be set to -INF or +INF for individual variables to disable bounding them. Complex boundary conditions can be handled in the performance function by returning INF whenever the input variables fall outside of range, though an initial conditions array should be used if this strategy is selected.