J. J. Merelo-Guervós > Algorithm-MasterMind-v0.4.4 > Algorithm::MasterMind::EvoRank

Algorithm-MasterMind-v0.4.4.tar.gz

Dependencies

Annotate this POD

# CPAN RT

 Open 0
View/Report Bugs
Module Version: 1.015   Source   Latest Release: Algorithm-MasterMind-v0.4.5

# NAME

Algorithm::MasterMind::EvoRank - Evolutionary algorithm with the partition method and ranked fitness, prepared for GECCO 2010

# SYNOPSIS

`    use Algorithm::MasterMind::Evolutionary_Partitions;`

# DESCRIPTION

The partition method was introduced in a 2010 paper, and then changed by Runarsson and Merelo to incorporate it in the genetic search. It was prepared for a conference paper, this one:

```  @INPROCEEDINGS{mm:cig,
author={Merelo, J.J. and Mora, A.M. and Runarsson, T.P. and Cotta, C.},
booktitle={Computational Intelligence and Games (CIG), 2010 IEEE Symposium on},
title={Assessing efficiency of different evolutionary strategies playing MasterMind},
year=2010,
month={August},
pages={38--45},
keywords={MasterMind player;constrained optimization problem;evolutionary algorithm;evolutionary strategy;fitness function;computer games;evolutionary computation;},
doi={10.1109/ITW.2010.5593373}
}```

This method is the evolutionary equivalent of Algorithm::MasterMind::Partitition::Most, using the number of non-null partitions to score consistent combinations, while using distance-to-consistency to score non-consistent.

# INTERFACE

## initialize

Initializes the genetic part of the algorithm

## issue_next()

Issues the next combination, using this method. Every generation runs an evolutionary algorithm to compute the next string.

## compute_fitness()

Processes "raw" fitness to assign fitness once consistency and/or distance to it is known. It's lineally scaled to make the lowest combination have a fitness equal to 1, which is needed just in case the selection method uses roulette wheel (which it does).

# AUTHOR

JJ Merelo `<jj@merelo.net>`

Copyright (c) 2009, 2010 JJ Merelo `<jj@merelo.net>`. All rights reserved.