gsl/simplex.py

The IMP::gsl::Simplex optimizer is a local optimizer that does not require derivatives.

## \example gsl/simplex.py
# The IMP::gsl::Simplex optimizer is a local optimizer that does not
# require derivatives.

import IMP
import IMP.test
import IMP.core
import IMP.gsl
import IMP.algebra
import sys

IMP.setup_from_argv(sys.argv, "simplex")

m = IMP.Model()

# create two particles to optimize
d0 = IMP.core.XYZ.setup_particle(
    IMP.Particle(m), IMP.algebra.Vector3D(0, 0, 0))
d1 = IMP.core.XYZ.setup_particle(
    IMP.Particle(m), IMP.algebra.Vector3D(3, 4, 5))
d0.set_coordinates_are_optimized(True)
d1.set_coordinates_are_optimized(True)

# restrain their distance to be 1
dist = IMP.core.DistanceRestraint(m, IMP.core.Harmonic(1, 1), d0, d1)

opt = IMP.gsl.Simplex(m)
opt.set_scoring_function([dist])
opt.set_minimum_size(.000001)
opt.set_initial_length(1)

# probably more steps than are needed
e = opt.optimize(1000000)
print(IMP.core.get_distance(d0, d1))