Table of contents

General

• A Simple Genetic Algorithm
• Defining GA's the easy way
• Evolution Strategy
• Genetic Mastermind
• Optimizing a vector quantizer
• Portability issues
  • Abstract Data Types
    • EO class
    • Fitness
    • Genetic operators
    • Selection and reproduction algorithms
  • EO Generic Classes
    • Generic genetic operators
    • Replacement algorithms
  • EO Specific Classes
    • EO Bidimemsional Operators
    • Operators for Bidimensional EOs
    • Operators for floating-point EOs
  • Utility classes
• Testing EO's chromosomes and operators
• Testing EO's chromosomes and operators
• Testing bidimensional chromosomes

Classes

• AgeStat class to get age statistics from a population
• Arg An Arg repesents an argument that can be passed to a program in the command line
• BinOp Binary genetic operator: subclasses EOOp, and defines basically the operator() with two operands
• CharCircMutate These are the genetic operators
• CoincidenceTerm Termination object for Genetic MasterMind
  Terminates when the number of rules met coincides with the number of rules so far, or if _maxCounter steps have happened without change
• EO EO: Base class for the whole hierarchy T is the type it will be instantiated with; this type must have, at least, a copy ctor, assignment operators,
• EOAgeGA Simple algorithm for evolutionary computation It is not exactly a Simple GA, as defined by Goldberg, it is just a simple implementation of a Genetic Algorithm
• EOAlgo Algorithm that performs the evolutionary computation Might be genetic or not
• EOBase Base class for chromosomes, it's a very basic class, which does not require anything but a fitness and an ID
• EOBidim Bidimensional Chromosomes
• EOBidimEvaluator Evaluator takes a vector of EOs and evaluates its fitness returning void
• EOBidimMutation Every gen is mutated adding or substracting a percentage multipliying by a value between 0 a 1
• EOBin Binary chromosome for EO; this is the classical bitstring chromosome described by Holland, Goldberg and others
• EOBinFactory Factory method
• EOBitFlip EOBitFlip: operator for binary chromosomes Changes one bit in the chromosome every time that is aplied
• EOBreeder Takes those on the selection list and creates a list of new individuals Destroys the genetic pool So many instances are necessary to be able to use EOops EOT is a EO-type class
• EOCrowding Each one of the new population substitutes the most similar of the old population
• EODup Dup or duplicate: duplicates a gene in a chromosome
• EOEasyGA Simple algorithm for evolutionary computation It is not exactly a Simple GA, as defined by Goldberg, it is just a simple implementation of a Genetic Algorithm
• EOElimAll Selects those who are going to reproduce, eliminates those which are not fit
• EOElimWorst Eliminates the worst of the populaton
• EOES Simple algorithm for evolutionary computation It is a wrapper for EasyGA, but lets the user only the decisions of some constants
• EOESChrom Basic chromosome for evolution strategies (ES), as defined by Rechenberg and Schwefel
• EOESMutate ES-style mutation: each gene in the chromosome is added its own sigma
  Obviously, valid only for EOESChrom
• EOEvaluator Evaluator takes a vector of EOs and evaluates its fitness returning void
• EOFactory EO Factory
• EOFitness Abstract data type for fitness objects.
• EOFitTerm Fitness termination: terminates after a the difference between the fitness of the best individual and a maximum fitness to achieve is less than certain number called epsilon
• EOFloatArrayMutateByPercentage Every gen is mutated adding or substracting a percentage multipliying by a value between 0 a 1
• EOFloatArrayXoverAverage Average crossover for arrays of floats
• EOFloatIntermXOver IntermXOver computes an arithmetic average of the two chromosomes
• EOGeneFactory Creates random EOs under request
• EOGenTerm Generational termination: terminates after a number of generations
• EOIntArrayMutate Mutation for integers: adds one
• EOKill Kill eliminates a gen in a chromosome
• EOLinearFreezer Implememtation of a linear freezer
• EOLottery EOLottery: an selection operator, which get the fitness from every chrommosome (fitness must have a standard typecast to float), and them creates a new population with fitness-proportionate selection
• EOMutation EOMutation: operator for binary chromosomes Mutation change for all bits in the chromosome
• EOOp Abstract data types for EO operators
• EOPop Subpopulation: it is used to move parts of population from one algorithm to another and one population to another
• EORandom EORandom: operator for binary chromosomes random change for all bits in the chromosome
• EORandomBreed Takes those on the selection list and creates a list of new individuals Destroys the genetic pool
• EORandomBreedLog Exactly as RandomBreed, except that uses factories
• EORandomSelect EORandomSelect: an selection operator, which selects random chromosomes
• EORank Takes those on the selection list and creates a list of new individuals Destroys the genetic pool
• EOReplace Abstract data type for Replacement algorithms, which will be implemented elsewhere
• EOSelector Selects those who are going to reproduce, eliminates those which are not fit
• EOSGA Simple algorithm for evolutionary computation It is a wrapper for EasyGA, but lets the user only the decisions of some constants
• EOSimAnnAlg Implementation of a Simulated annealing algorithm using EO
• EOSimAnnFreezer Abstrac implementation of a freezer
• EOStat class to get statistic from populations
• EOSteadyState Selects those who are going to reproduce, eliminates those which are not fit
• EOString Adaptor that turns an STL string into an EO
• EOTerm Termination condition for the genetic algorithm Takes the population as input, returns true for continue, false for termination
• EOTournament Selects those who are going to reproduce, eliminates those which are not fit using Tournament selection: a subset of the population of size tournamentSize is chosen, and the best is selected for the population
• EOTranspose Transposition operator: interchanges the position of two genes of an EO
• EOUniformCrossover EOUniformCrossover: operator for binary chromosomes implementation of uniform crossover for EO swaps ranges of bits between the parents
• EOVector Adaptor that turns an STL vector into an EO with the same gene type as the type with which the vector has been instantiated
• EOVectorFactory Factory method
• EOView class EOView: to decode binary chromosomes
• EOXOver2 2-point crossover: takes the genes in the central section of two EOs and interchanges it
• EvalAll Particular instantiation of the EOEvaluator class It takes each member in the population, and evaluates it, applying the evaluation function itīs been initialized with
• FitStat a simpler class to get fitness statistics from a population
• Flag A flag is a type of param that gets enabled if appears in the command line or disabled if it doesn't
• GeneCrossover GeneCrossover
• Grapher This is a decorator for the EOEvaluator class, as defined in the Gamma et al book
• InverOver Michalewicz's inver-over operator
• LenStat a simpler class to get length statistics from a population
• MonOp MonOp is the monary operator: genetic operator that takes only one EO
• NaryOp NaryOp is the N-ary operator: genetic operator that takes several EOs
• NxOver NxOver: multi point crossover
• Param A param is a type of argument that must be followed by its associated value in the command line or takes a default value if doesn't appear in the command line
• Parser Parses the command line
• Random Random Generator Class
• ScanXOver ScanXOver is the N-ary operator: genetic operator that takes several EOs
• TSPFitness Fitness for travellins salesman problem
• UException This class manages exceptions
• Uniform This class generates uniform random values of any type Uses the static class Random
• urEO Base class for chromosomes, it's a very basic class, which does not require anything but a fitness and an ID; and it know how to evaluate fitness itself

Functions, Macros

• () Generates a random T
• ::rChar Generates a random unsigned char
• createInitialPop Function that creates the initial population
• createInitialPop Function that creates the initial population
• evaluate
• getParams Parses the programīs command line, by creating a parsing object
• getParams Parses the programīs command line, by creating a parsing object
• getParams Parses the programīs command line, by creating a parsing object
• operator < Tricky operator to avoid errors in some VC++ systems
• operator < This operator is defined to avoid errors in some systems
• operator << Standard output
• operator << Standard output
• operator <<
• operator == Tricky operator to avoid errors in some VC++ systems
• operator == This operator is defined to avoid errors in some systems
• operator<<
• operator<<
• randomGen random generation of strings
• Uniform Constructor

Enums, Unions, Structs

• Arity
• EOESGene Each gene in an Evolution Strategies is composed of a value plus an standard deviation, sigma, used for mutation
• EOEvalFunc Evaluate: takes one EO and sets its "fitness" property returning this fitness also
• EvalRules Evaluation function object: it counts how many mastermind rules have been met
• MinOnes Evaluation Function to minimize the number of 1's in an EO
• NumOnes Implementation of a function evaluation object
• SortEO sort class for EOBase
• SortEOp for sorting EO Ops
• SortEOpPt For sorting EOOp pointers, can be used later on