Added more files from before

Author: fork-while-fork

ga-image-generation/GA_image_gen.py

@@ -0,0 +1,64 @@
+from pyevolve import G1DList
+from pyevolve import GSimpleGA
+from pyevolve import Consts
+from pyevolve import Selectors
+from pyevolve import Initializators
+from pyevolve import Mutators
+from pyevolve import Crossovers
+
+from PIL import Image
+
+image = Image.open("lena64.bmp")
+image_pixels = []
+for i in range(image.size[0]): 
+    for j in range(image.size[1]):
+        image_pixels.append(image.getpixel((i,j)))
+
+n = 0
+o = 0
+def callback(ga_engine):
+    global n
+    global o
+    if o % 10 == 0:
+        generation = ga_engine.getCurrentGeneration()
+        current_best = ga_engine.bestIndividual()
+        im = Image.new(image.mode, image.size)
+        i = 0
+        j = 0
+        for pixel in current_best:
+            im.putpixel((i, j), pixel)
+            j += 1
+            if j == image.size[1]:
+                i += 1
+                j = 0
+        im.save("./images/image%d.bmp" % n)
+        n += 1
+    o += 1
+    return False
+
+def fitness(genome):
+    return sum(abs(a-b) for a, b in zip(genome, image_pixels))
+
+def main():
+    genome = G1DList.G1DList(len(image_pixels))
+    genome.setParams(rangemin=min(image_pixels),
+                     rangemax=max(image_pixels),
+                     bestrawscore=0.00)
+ 
+    genome.initializator.set(Initializators.G1DListInitializatorInteger)
+    genome.mutator.set(Mutators.G1DListMutatorIntegerRange)
+    genome.evaluator.set(fitness)
+ 
+    ga = GSimpleGA.GSimpleGA(genome)
+    ga.stepCallback.set(callback)
+    ga.setMinimax(Consts.minimaxType["minimize"])
+    ga.terminationCriteria.set(GSimpleGA.RawScoreCriteria)
+    ga.selector.set(Selectors.GRankSelector)
+    ga.setPopulationSize(20)
+    ga.setMutationRate(0.005)
+    ga.setCrossoverRate(.75)
+    ga.setGenerations(5000)
+    ga.evolve(freq_stats=100)
+
+if __name__ == "__main__":
+    main()

ga-image-generation/bwkey.bmp

@@ -0,0 +1,2 @@
+The code in this directory (and all subdirectories) models the algorithm 
+described in the paper (ga-paper.pdf) that resides in this directory.