ec.app.gpsemantics.semantic.params Maven / Gradle / Ivy

Show more of this group Show more artifacts with this name
Show all versions of ecj Show documentation

ECJ, A Java-based Evolutionary Computation Research System. ECJ is a research EC system written in Java. It was designed to be highly flexible, with nearly all classes (and all of their settings) dynamically determined at runtime by a user-provided parameter file. All structures in the system are arranged to be easily modifiable. Even so, the system was designed with an eye toward efficiency. ECJ is developed at George Mason University's ECLab Evolutionary Computation Laboratory. The software has nothing to do with its initials' namesake, Evolutionary Computation Journal. ECJ's sister project is MASON, a multi-agent simulation system which dovetails with ECJ nicely.

The newest version!

# Copyright 2012 by James McDermott
# Licensed under the Academic Free License version 3.0
# See the file "LICENSE" for more information

parent.0 = ../../gp/koza/koza.params

# Function sets. Just one set, named gp.fs.0: it has 17 Xi variables
# and 17 Ni variables. Change the fs.0.size parameter to tune the
# problem.

gp.fs.size = 1
gp.fs.0.name = f0
gp.fs.0.size = 35
gp.fs.0.func.0 = ec.app.gpsemantics.func.SemanticJ
gp.fs.0.func.0.nc = nc2
gp.fs.0.func.1 = ec.app.gpsemantics.func.SemanticX0
gp.fs.0.func.1.nc = nc0
gp.fs.0.func.2 = ec.app.gpsemantics.func.SemanticN0
gp.fs.0.func.2.nc = nc0
gp.fs.0.func.3 = ec.app.gpsemantics.func.SemanticX1
gp.fs.0.func.3.nc = nc0
gp.fs.0.func.4 = ec.app.gpsemantics.func.SemanticN1
gp.fs.0.func.4.nc = nc0
gp.fs.0.func.5 = ec.app.gpsemantics.func.SemanticX2
gp.fs.0.func.5.nc = nc0
gp.fs.0.func.6 = ec.app.gpsemantics.func.SemanticN2
gp.fs.0.func.6.nc = nc0
gp.fs.0.func.7 = ec.app.gpsemantics.func.SemanticX3
gp.fs.0.func.7.nc = nc0
gp.fs.0.func.8 = ec.app.gpsemantics.func.SemanticN3
gp.fs.0.func.8.nc = nc0
gp.fs.0.func.9 = ec.app.gpsemantics.func.SemanticX4
gp.fs.0.func.9.nc = nc0
gp.fs.0.func.10 = ec.app.gpsemantics.func.SemanticN4
gp.fs.0.func.10.nc = nc0
gp.fs.0.func.11 = ec.app.gpsemantics.func.SemanticX5
gp.fs.0.func.11.nc = nc0
gp.fs.0.func.12 = ec.app.gpsemantics.func.SemanticN5
gp.fs.0.func.12.nc = nc0
gp.fs.0.func.13 = ec.app.gpsemantics.func.SemanticX6
gp.fs.0.func.13.nc = nc0
gp.fs.0.func.14 = ec.app.gpsemantics.func.SemanticN6
gp.fs.0.func.14.nc = nc0
gp.fs.0.func.15 = ec.app.gpsemantics.func.SemanticX7
gp.fs.0.func.15.nc = nc0
gp.fs.0.func.16 = ec.app.gpsemantics.func.SemanticN7
gp.fs.0.func.16.nc = nc0
gp.fs.0.func.17 = ec.app.gpsemantics.func.SemanticX8
gp.fs.0.func.17.nc = nc0
gp.fs.0.func.18 = ec.app.gpsemantics.func.SemanticN8
gp.fs.0.func.18.nc = nc0
gp.fs.0.func.19 = ec.app.gpsemantics.func.SemanticX9
gp.fs.0.func.19.nc = nc0
gp.fs.0.func.20 = ec.app.gpsemantics.func.SemanticN9
gp.fs.0.func.20.nc = nc0
gp.fs.0.func.21 = ec.app.gpsemantics.func.SemanticX10
gp.fs.0.func.21.nc = nc0
gp.fs.0.func.22 = ec.app.gpsemantics.func.SemanticN10
gp.fs.0.func.22.nc = nc0
gp.fs.0.func.23 = ec.app.gpsemantics.func.SemanticX11
gp.fs.0.func.23.nc = nc0
gp.fs.0.func.24 = ec.app.gpsemantics.func.SemanticN11
gp.fs.0.func.24.nc = nc0
gp.fs.0.func.25 = ec.app.gpsemantics.func.SemanticX12
gp.fs.0.func.25.nc = nc0
gp.fs.0.func.26 = ec.app.gpsemantics.func.SemanticN12
gp.fs.0.func.26.nc = nc0
gp.fs.0.func.27 = ec.app.gpsemantics.func.SemanticX13
gp.fs.0.func.27.nc = nc0
gp.fs.0.func.28 = ec.app.gpsemantics.func.SemanticN13
gp.fs.0.func.28.nc = nc0
gp.fs.0.func.29 = ec.app.gpsemantics.func.SemanticX14
gp.fs.0.func.29.nc = nc0
gp.fs.0.func.30 = ec.app.gpsemantics.func.SemanticN14
gp.fs.0.func.30.nc = nc0
gp.fs.0.func.31 = ec.app.gpsemantics.func.SemanticX15
gp.fs.0.func.31.nc = nc0
gp.fs.0.func.32 = ec.app.gpsemantics.func.SemanticN15
gp.fs.0.func.32.nc = nc0
gp.fs.0.func.33 = ec.app.gpsemantics.func.SemanticX16
gp.fs.0.func.33.nc = nc0
gp.fs.0.func.34 = ec.app.gpsemantics.func.SemanticN16
gp.fs.0.func.34.nc = nc0


# problem specification
eval.problem = ec.app.gpsemantics.Semantic
eval.problem.data = ec.gp.GPData


# parameters are as specified in Goldberg and O'Reilly:

# Parameters are specified as follows by Goldberg and O'Reilly: "Our
# runs use a problem sizeof 16 and the uniform fitness structure of
# the problem implies that the maximum adjusted fitness is
# 16+1=17. The height of trees in the initial population is 2. Maximum
# tree height is 15. The population size is 100.  Tournament selection
# is used. Runs proceed for a maximum of 50 generations.  Because the
# distribution of leaf to non-leaf primitives in the primitive set is
# skewed toward leafs, random tree creation is modified. First a
# equi-probable random choice is made between using a non-leaf or
# leaf. Then, if the choice is a leaf, one of the 32 leafs is randomly
# chosen.  In our simple version of GP, we have GP crossover available
# with the 90:10 biased node to leaf crossover point selection and
# with non-biased crossover point selection."

# koza.params brings in 90/10 crossover, replication 0.1, ramped half
# and half initialisation

# override a few koza.params values

# note that we don't have an initialiser that does exactly what
# Goldberg and O'Reilly say, ie choosing terminal v non-terminal
# first, then choosing which terminal if needed. But with ramped
# half-and-half in [2, 2], setting growp to 0.1 at least assures us we
# don't get mostly trivial trees.

gp.koza.half.growp = 0.1
gp.koza.half.min-depth = 2
gp.koza.half.max-depth = 2

pop.subpop.0.species.fitness = ec.simple.SimpleFitness

gp.koza.xover.maxdepth = 15
gp.koza.mutate.maxdepth = 15

generations = 50
pop.subpop.0.size = 100

eval.problem.problem_name = Order
# alternative, commented out:
# eval.problem.problem_name = Majority