A parallel branch-and-bound engine for Python.

Overview

pybnb

PyPI-Status PyPI-Versions PyPI-Downloads

Testing-Status Coverage-Status Documentation-Status

Codacy-Grade Mypy Black

A parallel branch-and-bound engine for Python. (https://pybnb.readthedocs.io)

This software is copyright (c) by Gabriel A. Hackebeil ([email protected]).

This software is released under the MIT software license. This license, including disclaimer, is available in the 'LICENSE' file.

Quick Start

Define a problem:

# simple.py

import pybnb
class Simple(pybnb.Problem):
    def __init__(self):
        self.bounds = [0.0,1.0]
    def sense(self):
        return pybnb.minimize
    def objective(self):
        return round(self.bounds[1] - self.bounds[0], 3)
    def bound(self):
        return -(self.bounds[1] - self.bounds[0])**2
    def save_state(self, node):
        node.state = self.bounds
    def load_state(self, node):
        self.bounds = node.state
    def branch(self):
        L, U = self.bounds
        mid = 0.5 * (L + U)
        for l,u in [(L,mid), (mid,U)]:
            child = pybnb.Node()
            child.state = (l,u)
            yield child

Write a solve script:

# solve_simple.py

import simple
problem = simple.Simple()
results = pybnb.solve(problem,
                      absolute_gap=1e-9)

Run the script:

$ mpirun -np 4 python solve_simple.py

Using non-default solver options:
 - absolute_gap: 1e-09 (default: 0)

Starting branch & bound solve:
 - dispatcher pid: 34902 (Ozymandias.local)
 - worker processes: 3
---------------------------------------------------------------------------------------------------------------------------
         Nodes        |                      Objective Bounds                        |              Work
      Expl    Unexpl  |      Incumbent           Bound     Rel. Gap         Abs. Gap | Time (s)  Nodes/Sec Imbalance   Idle
         0         1  |            inf            -inf          inf%             inf |      0.0       0.00     0.00%      0
*        1         2  |              1              -1  200.0000000%               2 |      0.0    1226.99   300.00%      1
*        2         3  |            0.5              -1  150.0000000%             1.5 |      0.0    2966.04   150.00%      0
*        4         5  |           0.25           -0.25   50.0000000%             0.5 |      0.0    8081.95    75.00%      0
*        8         9  |          0.125         -0.0625   18.7500000%          0.1875 |      0.0   12566.90    37.50%      0
      Expl    Unexpl  |      Incumbent           Bound     Rel. Gap         Abs. Gap | Time (s)  Nodes/Sec Imbalance   Idle
*       16        17  |          0.062       -0.015625    7.7625000%        0.077625 |      0.0   15352.74    18.75%      0
*       32        33  |          0.031     -0.00390625    3.4906250%      0.03490625 |      0.0   15981.49    18.75%      0
*       64        65  |          0.016   -0.0009765625    1.6976563%    0.0169765625 |      0.0   18740.68    18.75%      0
*      128       129  |          0.008   -0.0002441406    0.8244141%  0.008244140625 |      0.0   21573.51    11.72%      0
*      256       257  |          0.004   -6.103516e-05    0.4061035%  0.004061035156 |      0.0   22166.96     8.20%      0
      Expl    Unexpl  |      Incumbent           Bound     Rel. Gap         Abs. Gap | Time (s)  Nodes/Sec Imbalance   Idle
*      512       513  |          0.002   -1.525879e-05    0.2015259%  0.002015258789 |      0.0   21177.00     5.86%      0
*     1024      1025  |          0.001   -3.814697e-06    0.1003815%  0.001003814697 |      0.1   19978.42     9.38%      0
*     2048      2049  |              0   -9.536743e-07    0.0000954% 9.536743164e-07 |      0.1   21606.45     5.42%      0
     24029     24030  |              0   -1.490116e-08    0.0000015% 1.490116119e-08 |      1.1   21961.03     5.98%      0
     46159     46160  |              0    -3.72529e-09    0.0000004% 3.725290298e-09 |      2.1   22120.75     5.73%      0
      Expl    Unexpl  |      Incumbent           Bound     Rel. Gap         Abs. Gap | Time (s)  Nodes/Sec Imbalance   Idle
     65537     65538  |              0   -9.313226e-10    0.0000001% 9.313225746e-10 |      3.0   22459.50     6.20%      0
---------------------------------------------------------------------------------------------------------------------------

Absolute optimality tolerance met
Optimal solution found!

solver results:
 - solution_status: optimal
 - termination_condition: optimality
 - objective: 0
 - bound: -9.313226e-10
 - absolute_gap: 9.313226e-10
 - relative_gap: 9.313226e-10
 - nodes: 65537
 - wall_time: 2.96 s
 - best_node: Node(objective=0)

Number of Workers:        3
Load Imbalance:       6.20%
 - min: 21355 (proc rank=3)
 - max: 22710 (proc rank=1)
Average Worker Timing:
 - queue:      80.78% [avg time: 109.6 us, count: 65537]
 - load_state:  0.44% [avg time: 596.1 ns, count: 65537]
 - bound:       0.59% [avg time: 796.1 ns, count: 65537]
 - objective:   3.52% [avg time:   4.7 us, count: 65537]
 - branch:      3.36% [avg time:   4.6 us, count: 65537]
 - other:      11.31% [avg time:  15.3 us, count: 65537]
Comments
  • Redesigning Node Serialization

    Redesigning Node Serialization

    This is a major rewrite of the serialization strategy for nodes.

    Improvements include:

    • much faster serial performance (especially with pypy)
    • removed numpy as a dependency
    • any pickle-able object can be assigned to node.state (no longer needs to be a numpy float array)
      • old way:
      def save_state(self, node):
          node.resize(2)
          node.state[:] = (self.lb, self.ub)
      def load_state(self, node):
          self.lb, self.ub = node.state[:]
      
      • new way:
      def save_state(self, node):
          node.state = (self.lb, self.ub)
      def load_state(self, node):
          self.lb, self.ub = node.state
      
    • lexicographic node priorities can now be used (fixes #4):
      • example: solver.solve(problem, queue_strategy=('bound','objective'))
    • added a config object to allow users to adjust serialization settings (e.g., use dill instead of pickle)
    opened by ghackebeil 6
  • Getting the answers & path out?

    Getting the answers & path out?

    Hey, cool library!

    Maybe I'm missing something here, but it doesn't seem like there's any way to get the answers or the path to the answers out? SolverResults doesn't have the best node and there's no bookeeping to keep the chain of nodes?

    opened by MisterTea 4
  • Access solver attributes within problem

    Access solver attributes within problem

    Hi ! Thank you for the development of this package ! I was wondering if there was a way to call the solver attributes inside the problem methods. For instance, I would like to have access to the best objective within the bound method in the following way :

    class Problem(pybnb.Problem):
        ...
        def bound(self):
            best_objective = self.retrieve_solver_attribute("best_objective")
            return do_bounding_using_best_objective()
        ...
    

    Thank you for your answer, Théo

    opened by TheoGuyard 2
  • Obtain optimal assignment

    Obtain optimal assignment

    Hi Gabriel

    Thanks for a great project!

    I am struggling to find a way to obtain the variables' assignments once I have solved the example knapsack problem and obtained an optimal solution. Somehow I cannot access the decision variables' values leading to the obtained optimal objective value.

    I am assuming there must be a way to obtain the values of each integer variables in the results object, but I cannot figure out how. I have gone through the documentation, but can't seem to find anything. Any help? :)

    opened by alexberndt 2
  • Option to disable queue bound tracking

    Option to disable queue bound tracking

    This PR adds a solver option, track_bound, that allows one to disable functionality that tracks the global queue bound until the solve terminates. This can significantly reduce the overhead for certain queue implementations (except for worst-bound first, which tracks the bound for free).

    opened by ghackebeil 1
  • [WIP] Drop Support for Python 2.x

    [WIP] Drop Support for Python 2.x

    This PR tracks changes that can be made when support for Python 2 is dropped at the end of 2019. So far, changes are almost entirely cosmetic:

    • drop enum34 and six dependency
    • replace class Name(object): with class Name:
    • drop 2 or 3 lines that dealt with bytes <-> str casting
    • use keyword-only arguments

    Type annotations are not included in this PR as they are not entirely supported in Python 3.5. In particular, x: <type> = <value> is only supported outside function declarations in Python 3.6+.

    opened by ghackebeil 1
  • Issue when bound() and objective() methods return numpy types (parallel only)

    Issue when bound() and objective() methods return numpy types (parallel only)

    If the objective/bound computation uses NumPy and happens to return, for instance, a numpy.float64 type, this creates issues when serializing these node attributes through marshal.dumps(...). When the dispatcher receives the node, the call to marshal.loads produces an object of type bytes rather than a number that can be used for computation.

    I think the best fix is to add some code on the worker side that appropriately casts the objective/bound value returned from the problem to an int or float built-in type (possibly do the same with the queue_priority attribute in case the user sets this).

    An alternative would be to use pickle to serialize these node attributes, but I believe this can be noticeable slower than marshal (double check this).

    bug 
    opened by ghackebeil 1
  • Knapsack example

    Knapsack example

    Can you include a knapsack implementation in your examples? I'm struggling a bit with the documentation. It would be helpful for newcomers to familiarize with your library.

    opened by ggaylord 1
  • Formalize nested solver implementation

    Formalize nested solver implementation

    Things to address:

    • [ ] Setting up the communicator tree and listeners could be made easier
    • [ ] Allow load balance and timing statistics to be sent up to the root solver
    TODO 
    opened by ghackebeil 1
  • Allow lexicographic node prioritization in the queue

    Allow lexicographic node prioritization in the queue

    This would require allowing a tuple to be assigned to Node.queue_priority rather than just a number. One could utilize this either by implementing a custom queue strategy, or by setting the queue_strategy option to a tuple of queue strategy names.

    TODO 
    opened by ghackebeil 0
  • Automatically save the best node

    Automatically save the best node

    It can currently be done by the user using optional Problem methods, but it should probably done by the solver.

    Possible implementation:

    • When a worker thinks it has a new best node, send it to the dispatcher on the next update.
    • The dispatcher sends the best node to workers on the final update (or possibly earlier with every update).
    • If load_solution=True (new solver options, default True), load_state will be called with this node rather than the root at the end of the solve. If load_solution=False, the node will be placed on the results object.
    TODO 
    opened by ghackebeil 0
  • Multihreading instead of multiprocessing

    Multihreading instead of multiprocessing

    Hi there, Will there ever be a multithreading support instead of the already implemented multiprocess version? This may be very useful if multiple nodes share a child, and the bound() function makes use of all parents infos.

    opened by DarioSimonato 1
Releases(0.6.2)
  • 0.6.2(Dec 10, 2019)

  • 0.6.1(Mar 30, 2019)

  • 0.6.0(Mar 12, 2019)

    This is a major release that changes default optimality and tolerance settings to make solver behavior more intuitive. It also includes a new option for reducing dispatcher overhead for non-default queue strategies.

    Source code(tar.gz)
    Source code(zip)
  • 0.5.2(Feb 13, 2019)

    This is a minor release that adds a configuration option to enable compression and fixes a serialization issues when the bound or objective Problem methods return NumPy scalar types.

    Source code(tar.gz)
    Source code(zip)
  • 0.5.1(Feb 11, 2019)

    This is a minor release that includes some improvements to the documentation as well as fixes for some edge case behavior for unbounded problems.

    Source code(tar.gz)
    Source code(zip)
  • 0.5.0(Feb 10, 2019)

    This is a major release that includes some backward incompatible changes. Major changes include:

    • automatically saving the best node (rather than just the best objective)
    • changing the signature of the branch method so that it has more clear semantics
    • removing / renaming some of the optional problem callbacks
    • adding built-in support for nested solver implementations
    Source code(tar.gz)
    Source code(zip)
  • 0.4.0(Jan 31, 2019)

  • 0.3.0(Jan 21, 2019)

    This is a major release that includes some backward incompatible changes. Major changes include

    • renaming of a few solver options
    • addition of options for early termination of a solve
    • new queue management strategies
    • improvements to the online documentation
    Source code(tar.gz)
    Source code(zip)
  • 0.2.9(Dec 3, 2018)

  • 0.2.8(Nov 26, 2018)

  • 0.2.7(Nov 26, 2018)

  • 0.2.6(Jul 13, 2018)

    This minor release includes changes that improve dispatcher performance. It also includes an additional queue strategy that prioritizes nodes with the best objective first.

    Source code(tar.gz)
    Source code(zip)
  • 0.2.5(May 30, 2018)

  • 0.2.4(May 26, 2018)

    This release fixes a bug in pyomo_tools that left uncollected messages on the communicator. See the CHANGELOG for additional details about this release.

    Source code(tar.gz)
    Source code(zip)
  • 0.2.3(May 22, 2018)

  • 0.2.2(May 22, 2018)

  • 0.2.0(May 22, 2018)

Owner
Gabriel Hackebeil
Gabriel Hackebeil
Arabic to Roman Converter in Python

Arabic-to-Roman-Converter Made together with https://github.com/goltaraya . Arabic to Roman Converter in Python. -Instructions: 1 - Make sure you have

Pedro Lucas Tomazeti Fernandes 6 Oct 28, 2021
CD for MachineLearnia

Codebase supporting my talk on CI/CD for MachineLearnia (Nov 12 2021) The dataset used is available here. The point of the talk is to demonstrate a si

0 Feb 23, 2022
How to use Microsoft Bing to search for leaks?

Installation In order to install the project, you need install its dependencies: $ pip3 install -r requirements.txt Add your Bing API key to bingKey.t

Ernestas Kardzys 2 Sep 21, 2022
TinyBar - Tiny MacOS menu bar utility to track price dynamics for assets on TinyMan.org

📃 About A simple MacOS menu bar app to display current coins from most popular

Al 8 Dec 23, 2022
An end-to-end encrypted chat

An end-to-end encrypted chat, that allows users to anonymously talk without ip logs, personal info, or need for registration.

Privalise 1 Nov 27, 2021
Fast STL (ASCII & Binary) importer for Blender

blender-fast-stl-importer Fast STL (ASCII & Binary) importer for Blender based on https://en.wikipedia.org/wiki/STL_(file_format) Technical notes: flo

Iyad Ahmed 7 Apr 17, 2022
Ramadhan countdown - Simple daily reminder about upcoming Ramadhan

Ramadhan Countdown Bot Simple bot for displaying daily reminder about Islamic pr

Abdurrahman Shofy Adianto 1 Feb 06, 2022
Procedural 3D data generation pipeline for architecture

Synthetic Dataset Generator Authors: Stanislava Fedorova Alberto Tono Meher Shashwat Nigam Jiayao Zhang Amirhossein Ahmadnia Cecilia bolognesi Dominik

Computational Design Institute 49 Nov 25, 2022
A VirtualBox manager with interactive mode

A VirtualBox manager with interactive mode

Luis Gerardo 1 Nov 21, 2021
little proyect to organize myself, but maybe can help someone else

TaskXT 0.1 Little proyect to organize myself, but maybe can help someone else Idea The main idea is to ogranize you work and stuff to do, but with onl

Gabriel Carmona 4 Oct 03, 2021
Simple script to match riders with drivers.

theBestPooler Simple script to match riders with drivers. It's a greedy, unoptimised search, so no guarantees that it works. It just seems to work (ve

Devansh 1 Nov 22, 2021
GWAS summary statistics files QC tool

SSrehab dependencies: python 3.8+ a GNU/Linux with bash v4 or 5. python packages in requirements.txt bcftools (only for prepare_dbSNPs) gz-sort (only

21 Nov 02, 2022
Hack CMU Go Local Project

GoLocal A submission for the annual HackCMU Hackathon. We built a website which connects shopper with local businesses. The goal is to drive consumers

2 Oct 02, 2021
Displays Christmas-themed ASCII art

Christmas Color Scripts Displays Christmas-themed ASCII art. This was mainly inspired by DistroTube's Shell Color Scripts Screenshots ASCII Shadow Tex

1 Aug 09, 2022
PIP VA TASHQI KUTUBXONALAR

39-dars PIP VA TASHQI KUTUBXONALAR KIRISH Avvalgi darsimizda Python bilan birga o'rnatluvchi, standart kutubxona va undagi ba'zi foydali modullar bila

Sayfiddin 3 Nov 25, 2021
A collection of Python library code for building Python applications.

Abseil Python Common Libraries This repository is a collection of Python library code for building Python applications. The code is collected from Goo

Abseil 2k Jan 07, 2023
A simple script that shows important photography times. written in python.

A simple script that shows important photography times. written in python.

John Evans 13 Oct 16, 2022
Bots in moderation and a game (for now)

Tutorial: come far funzionare il bot e durarlo per 24/7 (o quasi...) Ci sono 17 passi per seguire: Andare sul sito Replit https://replit.com/ Vedrete

ZacyKing 1 Dec 27, 2021
Our product DrLeaf which not only makes the work easier but also reduces the effort and expenditure of the farmer to identify the disease and its treatment methods.

Our product DrLeaf which not only makes the work easier but also reduces the effort and expenditure of the farmer to identify the disease and its treatment methods. We have to upload the image of an

Aniruddha Jana 2 Feb 02, 2022
(Pre-)compromise operations for MITRE CALDERA

(Pre-)compromise operations for CALDERA Extend your CALDERA operations over the entire adversary killchain. In contrast to MITRE's access plugin, cald

Diederik Bakker 3 Aug 22, 2022