The Sequential Simplex Approach to Optimisation

John Ogilvie, CECM

     Although Maple has for several years contained procedures for
constrained optimization according to linear programming, another
approach involving simplices is applicable to non-linear regression.
Traditionally the latter approach provides no estimates of
uncertainty of parameters, but E. Romero and I have extended a
procedure in Maple to encompass this capability.  I shall demonstrate
with animations the application of this approach to real and
simulated data.