Running an Optimization Problem

This guide will walk you through the steps to set up and run an optimization problem using the PyGlobalSearch package in Python. We will cover defining the objective function, setting variable bounds, configuring the optimization parameters, and executing the optimization process.

Setting Up the Environment

Make sure to install the PyGlobalSearch package. You can do this using pip:

pip install pyglobalsearch

You will also need numpy for handling arrays. If you don’t have it installed, you can install it using pip:

pip install numpy

Defining your Problem

First, we need to define our optimization problem. We will create a Problem instance using the provided function by pyglobalsearch. We will implement the objective method to define our objective function and the variable_bounds method to set the bounds for each variable.

import pyglobalsearch as gs
import numpy as np

def objective_function(x):
    # Define your objective function here
    return x[0]**2 + x[1]**2

def variable_bounds():
    # Define the bounds for each variable
    return [(-10, 10), (-10, 10)]

problem = gs.Problem(objective=objective_function, variable_bounds=variable_bounds)

In the example above, we define an objective function that computes the sum of squares of the variables and set the variable bounds to be between -10 and 10 for both variables.

Configuring the Optimization Parameters

After setting up the problem, we need to configure the optimization parameters. This includes setting the population size, number of iterations, and other algorithm-specific parameters. We can do this using the OQNLPParams class.

We recommend reading the Overview documentation to understand how the algorithm works.

params = gs.PyOQNLPParams(
 iterations=100,
 population_size=500,
 wait_cycle=10,
 threshold_factor=0.75,
 distance_factor=0.1,
)

Make sure to adjust the parameters according to your specific optimization problem and requirements.

Running the Optimization

Now that we have defined our problem and configured the parameters, we can create an OQNLP instance and run the optimization process.

result = gs.optimize(problem, params, local_solver="COBYLA", seed=0)
print(result)

Your complete code should look like this:

import pyglobalsearch as gs
import numpy as np

def objective_function(x):
    # Define your objective function here
    return x[0]**2 + x[1]**2

def variable_bounds():
    # Define the bounds for each variable
    return [(-10, 10), (-10, 10)]

problem = gs.Problem(objective=objective_function, variable_bounds=variable_bounds)
params = gs.PyOQNLPParams(
 iterations=100,
 population_size=500,
 wait_cycle=10,
 threshold_factor=0.75,
 distance_factor=0.1,
)

result = gs.optimize(problem, params, local_solver="COBYLA", seed=0)
print(result)