knapsack_python 0.1.3

Solves a variety of knapsack problems

knapsack_python: Solves a variety of knapsack problems

This package is a collection of solutions to various knapsack problems. In particular, it has solutions to:

• the 01 knapsack problem,

• the 01 multi-knapsack problem (MKP),

and potentially more in the future. A good introduction to these sorts of problems can be found on Wikipedia (here and here). Additionally, it contains functions I’ve found useful in my work. One such function is assign_all, which assigns all items to one or more knapsacks while trying to adhere as best as possible to the capacities of each knapsack. Most of the solutions are a direct translation of the solutions given in Silvano Martello and Paolo Toth excellent book *Knapsack Problems: Algorithms and Computer Implementations*.

The implementations of these solutions are all written in C++ and wrapped in Cython for use in Python.

Quickstart

Installation

• pip install knapsack_python

or

• git clone https://github.com/python_knapsack

• cd python_knapsack

• python setup.py install

Use

All functions live in the knapsack_python module.

Example

Dependencies

• numpy

TODOs

• More comprehensive documentation

• Implement other knapsack-related problems such as:

  • 0-1 Knapsack Problem (really a special case of MKP)

  • Multiple-Choice Knapsack Problem

  • Bounded/Unbounded Knapsack Problem

  • Change-Making Problem

  • Generalized Assignment Problem

0.1.0

• Initial release with implementations of:

  • The MTHM algorithm for solving the multi-knapsack problem

  • A function to force assignment of all items while trying to respect the knapsacks’ capacities as much as possible