MHLib.jl - A Toolbox for Metaheuristics and Hybrid Optimization Methods
This project is still in early development, any feedback is much appreciated!
MHLib.jl is a collection of modules, types, and functions in Julia 1.5+ supporting
the effective implementation of metaheuristics and certain hybrid optimization approaches
for solving primarily combinatorial optimization problems.
MHLib.jl emerged from the
mhlib and the older
mhlib to which it has certain similarities
but also many differences.
The main purpose of the library is to support rapid prototyping and teaching as well as efficient implementations due to Julia's highly effective just-in-time-compilation.
MHLib.jl is developed primarily by the
Algorithms and Complexity Group of TU Wien,
Vienna, Austria, since 2020.
- Günther Raidl (primarily responsible)
- Nikolaus Frohner
- Thomas Jatschka
- Fabio Oberweger
Major versions of
pymhlib can be installed from the Julia REPL via
] add MHLib
Development versions are available at https://github.com/ac-tuwien/MHLib.jl and can be installed via
] add https://github.com/ac-tuwien/MHLib.jl.git
MHLib.jl is still far behind the capabilities of the Python
The main module provides the following types for candidate solutions and various functions for them:
Solution: An abstract type that represents a candidate solution to an optimization problem.
VectorSolution: An abstract solution encoded by a vector of some user-provided type.
BoolVectorSolution: An abstract solution encoded by a boolean vector.
PermutationSolution: An abstract solution representing permutations of a fixed number of elements. _
SubsetVectorSolution: A solution that is an arbitrary cardinality subset of a given set represented in vector form. The front part represents the selected elements, the back part optionally the unselected ones.
Moreover, the main module provides:
git_version(): Function returning the abbreviated git version string of the current project.
settings: Global settings that can be defined independently per module in a distributed way, while values for these parameters can be provided as program arguments or in configuration files. Most
pymhlibmodules rely on this mechanism for their external parameters.
Scheduler: A an abstract framework for single trajectory metaheuristics that rely on iteratively applying certain methods to a current solution. Modules like
ALNSsextend this type towards more specific metaheuristics.
GVNSs: A framework for local search, iterated local search, (general) variable neighborhood search, GRASP, etc.
ALNS: A framework for adaptive large neighborhood search (ALNS).
For demonstration purposes, simple metaheuristic approaches are provided in the
subdirectory for the following well-known combinatorial optimization problems.
They can be started as shown in the respective sections of
It is recommended to take such a demo as template for solving your own problem.
OneMax: basic test problem in which the goal is to set all digits in a binary string to
GraphColoring: graph coloring problem based on
MAXSAT: maximum satisfiability problem based on
TSP: traveling salesperson problem based on
MKP: multi-constrained knapsack problem based on
MISP: maximum independent set problem based on