Liking cljdoc? Tell your friends :D


clojure.yml Clojars Project

clojure(script) genetic algorithms library


Add the following dependency to your project.clj:

[com.github.kongeor/chickn "0.1.105"]


com.github.kongeor/chickn {:mvn/version "0.1.105"}

if you are using deps.edn


In the following example we will try to solve one of the most trivial problems to understand the concepts of the library:

First we need to create a function that will randomize ones and zeros. This will be used for the initial population but also for our mutation function:

(def one-or-zero (fn [& _] (if (> (rand) 0.5) 1 0)))

Let's define a population size:

(def population-size 20)

The chromo-gen function is used to create the initial population:

(def chromo-gen #(repeatedly population-size one-or-zero))

Fitness is the function that assigns a score to each possible solution. In this case it's just the sum of all numbers. Fitness for solution [0 0 1 1 0] would be 2, for [0 0 1 1 1] 3 etc.

(defn fitness [xs]
  (apply + xs))

In some cases we may be able to define a function that determines if the problem is solved, which will allow us to avoid wasting iteration cycles when we have found the solution we are looking for. This is an optional key.

(defn solved? [_ {:keys [best-chromosome]}]
  (every? #(= 1 %) best-chromosome))

We need to customize the build-in rand-mutation operator and specify the mutation function, which is the same we used for initializing the population. This is suboptimal, we could have just flipped the bit here, but for this example it should be all right.

(def mutation-op
          {:type          :chickn.mutation/rand-mutation
           :rate          0.3
           :random-func   rand
           :mutation-func one-or-zero})

Chickn comes with a default config, but some customization is needed (chromo-gen and fitness have to be provided). Here we are also specifying the :solved? which was explained above. We are muting the :reporter to avoid getting prints, setting up our mutation operator, and setting that solutions with higher scores are preferred.

(def config (merge
                      {:chromo-gen chromo-gen
                       :fitness    fitness
                       :solved?    solved?
                       :reporter   util/noop
                       :mutation   mutation-op
                       :comparator higher-is-better}))

It's time to fire the process!

    (init-and-evolve config 100) :population)
=> {:solved? true, :iteration 9, :best-chromosome [1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1], :time 1}

We are dissocing the :population as it will include the entire population and the output can be quite verbose.

You can find the code for this example here src/chickn/examples/hello_world.cljc.


Examples namespace has a few code examples.

Using chickn from cljs for solving the Traveling Salesman Problem.

Evolduo is using Chickn for evolving musical phrases.

Project Status

Chickn should be considered alpha quality software.


Copyright © 2018-2023 Kostas Georgiadis

Distributed under the Eclipse Public License either version 1.0 or (at your option) any later version.

Can you improve this documentation?Edit on GitHub

cljdoc is a website building & hosting documentation for Clojure/Script libraries

× close