The Nelder-Mead method

The easy implementation of the Nelder-Mead method.

Description

The Nelder-Mead is the one of derivative-free optimization method. This method is called simplex method or ameba method.

Installation

pip install git+https://github.com/owruby/nelder_mead.git

Dependencies

• numpy

Usage

from nelder_mead import NelderMead


def sphere(x):
    return sum([t**2 for t in x])


def main():
    func = sphere
    params = {
        "x1": ["real", (-512, 512)],
        "x2": ["real", (-512, 512)],
    }

    nm = NelderMead(func, params)
    nm.minimize(n_iter=30)


if __name__ == "__main__":
    main()

References

• https://academic.oup.com/comjnl/article-abstract/7/4/308/354237/A-Simplex-Method-for-Function-Minimization
• http://epubs.siam.org/doi/book/10.1137/1.9780898718768