Skip to main content

A SciPy compatible super fast Python implementation for Particle Swarm Optimization.

Project description

A python implementation of Particle Swarm Optimization.

Introduction

PSOPy (pronounced “Soapy”) is a SciPy compatible super fast Python implementation for Particle Swarm Optimization. The codes are tested for standard optimization test functions (both constrained and unconstrained).

The library provides two implementations, one that mimics the interface to scipy.optimize.minimize and one that directly runs PSO. The SciPy compatible function is a wrapper over the direct implementation, and therefore may be slower in execution time, as the constraint and fitness functions are wrapped.

Installation

GitHub

To install this library from GitHub,

$ git clone https://github.com/jerrytheo/psopy.git
$ cd psopy
$ python setup.py install

In order to run the tests,

$ python setup.py test

PyPI

This library is available on the PyPI as psopy. If you have pip installed run,

$ pip install psopy

Examples

Unconstrained Optimization

Consider the problem of minimizing the Rosenbrock function, implemented as scipy.optimize.rosen using a swarm of 1000 particles.

>>> import numpy as np
>>> from psopy import minimize
>>> from scipy.optimize import rosen
>>> x0 = np.random.uniform(0, 2, (1000, 5))
>>> res = minimize(rosen, x0, options={'stable_iter': 50})
>>> res.x
array([1.00000003, 1.00000017, 1.00000034, 1.0000006 , 1.00000135])

Constrained Optimization

Next, we consider a minimization problem with several constraints. The intial positions for constrained optimization must adhere to the constraints imposed by the problem. This can be ensured using the provided function psopy.init_feasible. Note, there are several caveats regarding the use of this function. Consult its documentation for more information.

>>> # The objective function.
>>> fun = lambda x: (x[0] - 1)**2 + (x[1] - 2.5)**2
>>> # The constraints.
>>> cons = ({'type': 'ineq', 'fun': lambda x:  x[0] - 2 * x[1] + 2},
...         {'type': 'ineq', 'fun': lambda x: -x[0] - 2 * x[1] + 6},
...         {'type': 'ineq', 'fun': lambda x: -x[0] + 2 * x[1] + 2},
...         {'type': 'ineq', 'fun': lambda x: x[0]},
...         {'type': 'ineq', 'fun': lambda x: x[1]})
>>> from psopy import init_feasible
>>> x0 = init_feasible(cons, low=0., high=2., shape=(1000, 2))
>>> res = minimize(fun, x0, constrainsts=cons, options={
...     'g_rate': 1., 'l_rate': 1., 'max_velocity': 4., 'stable_iter': 50})
>>> res.x
array([ 1.39985398,  1.69992748])

Authors

License

Licensed under the BSD 3-Clause License.
Copyright 2018 Abhijit Theophilus, Snehanshu Saha, Suryoday Basak

References

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

psopy-0.2.4.tar.gz (14.2 kB view hashes)

Uploaded Source

Built Distribution

psopy-0.2.4-py3-none-any.whl (14.9 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page