API Reference for Enthought Tool Suite 3.0.1
COBYLA: A python interface to the COBYLA non-linear optimizer
COBYLA, a contrained optimization by linear approximation package developed by Michael J. D. Powell, is a non-linear derivative-free optimizer. To use it, you must provide a function to minimize. The function must take the list of coordinates where to evaluate the function; and it must return either a tuple, whose first element is the value of the function, and whose second argument is the values of the constraints (as a list of values); or None, to abort the minimization.
COBYLA will try to make all the values of the constraints positive. So if you want to input a constraint j such as x[i] <= MAX, set:
con[j] = MAX - x[i]
HUGE_VAL = (9.9999999999999997e+199 * 9.9999999999999997e+199)
MAXFUN = 1
MSG_EXIT = 1
MSG_INFO = 3
MSG_ITER = 2
MSG_NONE = 0
MSGS = { Name('MSG_NONE') : Const('No messages'), Name('MSG_EXIT') : Const('Exit reasons'), Name('MSG_ITER') : Const('Rho and Sigma changes'), Name('MSG_INFO') : Const('Informational messages') }
NORMAL = 0
RCSTRINGS = { Name('NORMAL') : Const('Normal return'), Name('MAXFUN') : Const('Max. number of function evaluations reach'), Name('ROUNDING') : Const('Rounding errors are becoming damaging'), Name('USERABORT') : Const('User requested end of minimization') }
ROUNDING = 2
USERABORT = 3
Minimize a function with variables subject to non-linear constraints
returns (rc, nfeval, x).
The function to minimize. Signature must be:
f, con = function(x)
where f is the value of the function and con is the value of the constraints (a list of m floats). If the function returns None, the minimization is aborted.
| Local name | Refers to |
|---|---|
| moduleCobyla | moduleCobyla |
© Copyright 2002-2008 Enthought, Inc.