Permalink
Browse files

small stuff - lets get the 0.2 ready for release some day

  • Loading branch information...
1 parent 7c53c4e commit a4c9950d75c50bedca5bcf418bc4ddc3ef52f577 @yadler committed Nov 18, 2010
Showing with 42 additions and 14 deletions.
  1. +2 −0 CHANGELOG
  2. +7 −4 README
  3. +5 −5 TODO
  4. +25 −2 docs/FEATURES
  5. +1 −1 examples/tutorial/lesson01_first_steps.rb
  6. +2 −2 examples/tutorial/lesson03_custom_operators.rb
View
@@ -1,5 +1,7 @@
=== 0.2.0
+* various bugfixes ;-)
+
* added SP_k problem to binary benchmark functions
* added EvoBench (benchmarking for evolutionary algorithms) module:
View
11 README
@@ -4,7 +4,7 @@ Website:: http://evosynth.rubyforge.org/
Project:: http://rubyforge.org/projects/evosynth/
Gem:: http://rubygems.org/gems/evosynth
Sourcecode:: http://gitorious.org/evosynth and http://github.com/yadler/EvoSynth
-Author:: Yves Adler (http://yadler.wordpress.com/, mailto:yves.adler@googlemail.com)
+Author:: Yves Adler (http://confusedbits.net/, mailto:yves.adler@googlemail.com)
Copyright:: Copyright (c) 2009, 2010 Yves Adler <yves.adler@googlemail.com>
License:: MIT (see LICENSE)
@@ -22,13 +22,17 @@ EvoSynth (Evolutionary Computation Synthesizer) is a framework for rapid develop
* classes for individuals, populations, algorithm configurations, genomes
* support for custom randomizer
* meta operators: proportional, sequentional and conditional combined operators
-* logging support with exporter to gnuplot, gruff and csv
+* logging support with exporter to gnuplot, html and csv
* many predefined fitness functions
-* benchmarking features:
+* benchmarking features (evobench module):
* diversity calculations (distance, entropy and subsequence)
* mean, median, variance calculations for array/population
* t-test to determine statistical significance
+ * Comparator, to compare the performance of two or more Evolvers
+ * TestRun: runs a evolver with a given configuration (n times) and collects the produced data
+ * Experiment class: run a experiment with a experimental plan and compare different parameters
+ * FullFactorialPlan : full factorial experimental plan
* most common evolutionary algorithms:
* hillclimber (single individual and population based)
@@ -59,4 +63,3 @@ EvoSynth (Evolutionary Computation Synthesizer) is a framework for rapid develop
* arithmetic crossover
* ordered recombination, partially mapped crossover, edge recombination
* global uniform crossover, global arithmetic crossover
-
View
10 TODO
@@ -1,12 +1,9 @@
============================== EvoSynth 0.2 ====================================
-Features:
-
- README:
- => features aktualisieren (siehe CHANGELOG)
- => CHANGELOG, README, TODO aufräumen, schick machen und mit commits vergleichen
=> Quellen irgendwo in der Readme oder in eigener Datei erwähnen (für Algorithmen)
- => Auf Diplomarbeitshintergrund verweisen?
+
+Features:
- gnuplot export:
=> bessere bzw. mehr optionen, templates
@@ -30,6 +27,9 @@ Features:
Features:
+ - Ruby 1.9.2:
+ => Randomizer krempel checken - da gibts Neuerungen in Ruby!
+
- Verbesserungen:
=> Evolver, Evaluator in ein Module transformieren:
-> statt Konstruktor #included verwenden!
View
@@ -3,6 +3,7 @@
* EvoSynth::Evaluator
* EvoSynth::Configuration
* EvoSynth::Population
+* EvoSynth::Logger
* EvoSynth::Individual
* EvoSynth::MaximizingIndividual
@@ -13,10 +14,11 @@
* EvoSynth.nrand
* EvoSynth.rand_bool
-* Include "Genotypes":
+* Genotypes:
* EvoSynth::ArrayGenome (can contain any datatype as items)
* EvoSynth::BinaryGenome (not optimized yet)
+* EvoSynth::Decoder
== Evolution algorithms
@@ -58,7 +60,6 @@
* EvoSynth::Selections::RouletteWheelSelection
* EvoSynth::Selections::RandomSelection
-
== Meta operators
* EvoSynth::MetaOperators::ProportionalCombinedOperator
@@ -109,3 +110,25 @@
* EvoSynth::Adjustments::PredifinedAdjustment
* EvoSynth::Adjustments::AdaptiveAdjustment
+
+== EvoBench: benchmarking of evolutionary algorithms
+
+* EvoSynth::EvoBench::Experiment
+* EvoSynth::EvoBench::RunResult
+* EvoSynth::EvoBench::FullFactorialPlan
+* EvoSynth::EvoBench::Comparator
+* EvoSynth::EvoBench also includes a couple of util functions (t-test, diversity, ...)
+
+== Export
+
+* EvoSynth::Export::ConsoleWriter
+* EvoSynth::Export::GnuPlot
+* EvoSynth::Export::CSV
+* EvoSynth::Export::HTML
+
+== Predefined benchmak (fitness-) functions
+
+* EvoSynth::Problems::BinaryBenchmarkFuntions
+* EvoSynth::Problems::FloatBenchmarkFuntions
+* EvoSynth::Problems::GraphColouring
+* EvoSynth::Problems::TSP
@@ -63,4 +63,4 @@ def calculate_fitness(individual)
puts "\nRunning hillclimber...\n"
result = evolver.run_while { |hc| hc.evaluator.called < 5001 }
-puts "\nIndividual after evolution: #{result}"
+puts "\nIndividual after evolution: #{result}"
@@ -42,7 +42,7 @@ def mutate(individual)
end
end
-# Next we create a stupid recombination. This time we've to overwrite the recombine methdo,
+# Next we create a stupid recombination. This time we've to overwrite the recombine method,
# which takes two individuals as arguments and returns a array of two individuals.
class MyRecombination
@@ -99,4 +99,4 @@ def calculate_fitness(individual)
puts "\nRunning genetic algrithm...\n"
result = evolver.run_while { |ga| ga.evaluator.called < 5001 }
-puts "\nPopulation after evolution: #{result}"
+puts "\nPopulation after evolution: #{result}"

0 comments on commit a4c9950

Please sign in to comment.