The Optimizer

The Optimizer is a python package that provides a collection of powerful optimization algorithms, including MOPSO (Multi-Objective Particle Swarm Optimization). The primary purpose of this package is to facilitate running optimization tasks using user-defined Python functions as the optimization target. The package is developed with the objectives of CMS and Patatrack in mind.

The Optimizer

Installation

To use The Optimizer package, you can install it using pip and the provided setup.py script.

Clone this repository
Navigate to the project directory
Install the package and its dependencies using pip:
```
pip install .
```

You can install a project in “editable” or “develop” mode while you’re working on it. When installed as editable, a project can be edited in-place without reinstallation:

python3 -m pip install -e .

Usage

Currently the package provides the mopso.py module that defines two classes: MOPSO and Particle. To use this module in your Python projects:

Import the required modules:
```
from optimizer import MOPSO
```

Define the objective function to be optimized.

def objective_function(x):
    return np.sin(x[0]) + np.sin(x[1])

Create the MOPSO object with the configuration of the algorithm

mopso = MOPSO(
    objective_functions=[objective_function],
    lower_bound=-5,
    upper_bound=5,
    num_particles=50,
    num_iterations=100
)

Run the optimization algorithm
```
pareto_front = mopso.optimize()
```

Objective Function

Depending on the optimization mode, the Objective function can be defined in two way:

In individual mode, the objective function is evaluated particle by particle at every iteration and is called as:

objective_funciton(self.position) # self is a Particle object

the argument of the optimization function is an array of elements corresponding to the parameters to optimize. The output is the fitness of the particle: the value(s) to minimize in order to solve the optimization problem.

See run_mopso.py for an example.

in global mode, the objective function is evaluated once per iteration and is called as:

objective_function([particle.position for particle in self.particles])

the argument of the optimization function is a list of arrays: nn array of parameters for each particle. The output is a list of fitnesses: the value(s) to minimize for each particle.

See run_mopso.py for an example.

Contributing

Contributions are welcome. If you want to contribute, please follow the Contribution guidelines.

License