Skip to content
This repository has been archived by the owner. It is now read-only.
High-performance evolutionary computation in Julia.
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.

Warning: Wallace is currently in an non-functional state, following changes from Julia 0.3 onwards. For now, users are encouraged to seek other solutions.


Wallace is a high-performance, easy-to-use evolutionary computation framework written in Julia, built for researchers, students, and software engineers alike. To simultaneously achieve performance and ease-of-use, Wallace utilises computational reflection to compile problem-specific data structures and algorithms at run-time, using a beautiful and compact domain-specific language.

Through its use of various programming techniques, and the thanks to the speed of Julia, Wallace is both the fastest evolutionary computation framework, beating ECJ and JCLEC, as well as the most expressive and concise, requiring fewer lines of code to write than DEAP. For these reasons, Wallace makes both a great framework for performing industrial-strength evolutionary computation on highly complex real-world problems, as well as teaching its concepts.


  • A compact, extensible domain-specific language.
  • Multi-processing and multi-threading.
  • Support for massively distributed computation.
  • Just-in-time compilation of problem-optimised algorithms.
  • Data logging, plotting and visualisation.
  • An interactive shell, with an embedded documentation browser (via Julia).
  • Modular structure makes it easy to extend and share components.

Algorithms, Representations, and Operations:

  • Vectors, lists, sets, trees, orderings, etc.
  • Grammatical evolution (efficient compilation to arbitrary languages).
  • Genetic programming: Koza, Strictly-Typed GP, Grammar-Guided GP, PushGP, Cartesian GP.
  • Multiple objective optimisation: NSGA, NSGA-II, SPEA, SPEA2, VEGA, and more.
  • Co-operative and competitive co-evolution approaches.
  • Island model support, for same and different species islands.


The latest stable release of Wallace is most easily installed using Julia's package manager within the REPL, as shown below.

$ julia

julia> Pkg.add("Wallace")

To install the latest developmental version of Wallace, open up the Julia REPL, and call the following:

julia> Pkg.remove("Wallace")
julia> Pkg.clone("git://")

This will remove any existing Wallace module from your Julia installation, before cloning the contents of the Wallace git repository into your local Julia modules directory.

Note: To avoid installation issues, ensure that the rest of your Julia packages are up-to-date via Pkg.update().

Documentation and Tutorials

For technical documentation and an extensive set of tutorials, aimed at a wide range of different audiences, visit the Wallace website at:

Up-to-date documentation and tutorials can be found at:


Below is the source code for the Max Ones benchmark problem provided in the examples package.

using Wallace

# Provide a definition for the algorithm.
def = algorithm.genetic() do alg
  alg.population = population.simple() do pop
    pop.size = 100

    # Species describes the fitness scheme and representation used by
    # individuals belonging to that species.
    pop.species = species.simple() do sp = fitness.scalar(Int)
      sp.representation = representation.bit_vector(100)

    # Multi-threading breeding.
    pop.breeder = breeder.simple() do br
      br.threads = 8
      br.selection = selection.tournament(2)
      br.mutation = mutation.bit_flip(1.0)
      br.crossover = crossover.one_point(0.1)

  # Evaluation function (split across 8 threads).
  alg.evaluator = evaluator.simple(; threads = 8) do scheme, genome
    assign(scheme, sum(genome))

  # Termination conditions.
  alg.termination = [criterion.generations(1000)]

# Compose the algorithm from its definition.
alg = compose!(def)

# Run the composed algorithm.


If you plan on using Wallace for your research, we encourage you to cite the paper below. Additionally, put in a merge request, and we will add your paper to the list of papers using Wallace.

  author = {Timperley, Christopher Steven and Stepney, Susan},
  title = {Wallace: An efficient generic evolutionary framework},
  booktitle={ECAL 15},
  organization={MIT Press}
You can’t perform that action at this time.