introtoga/chap1/comp1/simpleGA.rkt

#lang typed/racket

;; A simple genetic algorithm
;; Computer Exercise #1 from Intro to GA by Mitchell
;;
;; Uses:
;; - fitness-proportionate selection
;; - routlette-wheel sampling

; The population is stored as a single list of popsize*chromlen elements

(require math)

; population size and individual chromosome length
(define popsize 100)
(define chromlen 100)
; single-point crossover rate
(define p-cross 0.7)
; bitwise mutation rate
(define p-mutate 0.001)


; Fitness assessment is the number of 1's in the population member
(: fitness (-> (Listof Integer) Real))
(define (fitness member)
  (/ (sum member)
     (length member)))

; Flip a chromosome to its opposite
(: flip-chrom (-> Integer Integer))
(define (flip-chrom val)
  (if (eq? val 1) 0 1))


; mutate
; because of the way the population is stored and that mutations happen after
; offspring are produced, we can mutate the whole list together
(: mutate (-> (Listof Integer) (Listof Integer)))
(define (mutate pop)
  (for/list: : Listof Integer ([i : Index pop])
    ((λ (chrom)
       (if (<= (random) p-mutate)
           (flip-chrom i)
           i))
     i)))

; Initialize a random population.
;(: initpop (-> (Listof Integer)))
;(define initpop
;  (for/list ([i (in-range (* popsize chromlen))])
;    (random-integer 0 1)))