https://img.shields.io/pypi/v/codableopt.svg

codable-model-optimizer

Optimization problem meta-heuristics solver for easy modeling.

Installation

Use pip

$ pip install codableopt

Use setup.py

# Master branch
$ git clone https://github.com/recruit-tech/codable-model-optimizer
$ python3 setup.py install

Example Usage

Sample1

import numpy as np
from codableopt import *

# set problem
problem = Problem(is_max_problem=True)

# define variables
x = IntVariable(name='x', lower=np.double(0), upper=np.double(5))
y = DoubleVariable(name='y', lower=np.double(0.0), upper=None)
z = CategoryVariable(name='z', categories=['a', 'b', 'c'])


# define objective function
def objective_function(var_x, var_y, var_z, parameters):
    obj_value = parameters['coef_x'] * var_x + parameters['coef_y'] * var_y

    if var_z == 'a':
        obj_value += 10.0
    elif var_z == 'b':
        obj_value += 8.0
    else:
        # var_z == 'c'
        obj_value -= 3.0

    return obj_value


# set objective function and its arguments
problem += Objective(objective=objective_function,
                     args_map={'var_x': x,
                               'var_y': y,
                               'var_z': z,
                               'parameters': {'coef_x': -3.0, 'coef_y': 4.0}})

# define constraint
problem += 2 * x + 4 * y + 2 * (z == 'a') + 3 * (z == ('b', 'c')) <= 8
problem += 2 * x - y + 2 * (z == 'b') > 3

print(problem)

solver = OptSolver()

# generate optimization methods to be used within the solver
method = PenaltyAdjustmentMethod(steps=40000)

answer, is_feasible = solver.solve(problem, method)
print(f'answer:{answer}, answer_is_feasible:{is_feasible}')

Sample2

import random
from itertools import combinations

from codableopt import Problem, Objective, CategoryVariable, OptSolver, PenaltyAdjustmentMethod


# define distance generating function
def generate_distances(args_place_names):
    generated_distances = {}
    for point_to_point in combinations(['start'] + args_place_names, 2):
        distance_value = random.randint(20, 40)
        generated_distances[point_to_point] = distance_value
        generated_distances[tuple(reversed(point_to_point))] = distance_value
    for x in ['start'] + args_place_names:
        generated_distances[(x, x)] = 0

    return generated_distances


# generate TSP problem
PLACE_NUM = 30
destination_names = [f'destination_{no}' for no in range(PLACE_NUM)]
place_names = [f'P{no}' for no in range(PLACE_NUM)]
distances = generate_distances(place_names)
destinations = [CategoryVariable(name=destination_name, categories=place_names)
                for destination_name in destination_names]

# set problem
problem = Problem(is_max_problem=False)


# define objective function
def calc_distance(var_destinations, para_distances):
    return sum([para_distances[(x, y)] for x, y in zip(
        ['start'] + var_destinations, var_destinations + ['start'])])


# set objective function and its arguments
problem += Objective(objective=calc_distance,
                     args_map={'var_destinations': destinations, 'para_distances': distances})

# define constraint
# constraint formula that always reaches all points at least once
for place_name in place_names:
    problem += sum([(destination == place_name) for destination in destinations]) >= 1

# optimization implementation
solver = OptSolver(round_times=4, debug=True, debug_unit_step=1000)
method = PenaltyAdjustmentMethod(steps=10000, delta_to_update_penalty_rate=0.9)
answer, is_feasible = solver.solve(problem, method, n_jobs=-1)

print(f'answer_is_feasible:{is_feasible}')
root = ['start'] + [answer[root] for root in destination_names] + ['start']
print(f'root: {" -> ".join(root)}')

Release files for codableopt 0.1.2

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for codableopt 0.1.2
File	Size	Uploaded
codableopt-0.1.2.tar.gz	93.4 kB	Aug 3, 2022	Details

Built distribution (wheel)

Table of built distributions (wheels) for codableopt 0.1.2
File	Interpreter	ABI	Platform	Reset
codableopt-0.1.2-py3-none-any.whl 79.4 kB Aug 3, 2022	Python 3	none	any	Details

Total release size: 172.8 kB

Release files / codableopt-0.1.2.tar.gz

Download URL	codableopt-0.1.2.tar.gz
Size	93.4 kB
Tags	Source
SHA-256 checksum How to use checksums	`9277badb1d842410283d8e059f9da2b67022e0b688d1e6f36b1863fcca299571`
BLAKE2b-256 checksum How to use checksums	`35b824dfad4b3a6c5a40bd62ef28b213147be9f7ddac1466a8b772555b36f758`
Upload date	Aug 3, 2022
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/3.8.0 pkginfo/1.8.2 readme-renderer/34.0 requests/2.27.1 requests-toolbelt/0.9.1 urllib3/1.26.8 tqdm/4.63.0 importlib-metadata/4.11.3 keyring/23.5.0 rfc3986/2.0.0 colorama/0.4.4 CPython/3.8.12`

Release files / codableopt-0.1.2-py3-none-any.whl

Download URL	codableopt-0.1.2-py3-none-any.whl
Size	79.4 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`881788ecad6048db122e96e4e728b139acb0e7efc80039271aeb0f34f1e9a778`
BLAKE2b-256 checksum How to use checksums	`1286c5213cc2e0885848c7f0048788fe2f92f4505d774dd8ebcff018b38c3066`
Upload date	Aug 3, 2022
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/3.8.0 pkginfo/1.8.2 readme-renderer/34.0 requests/2.27.1 requests-toolbelt/0.9.1 urllib3/1.26.8 tqdm/4.63.0 importlib-metadata/4.11.3 keyring/23.5.0 rfc3986/2.0.0 colorama/0.4.4 CPython/3.8.12`