gga.cc

linux下的一个分组遗传算法

//
// gga.cc - Grouping Genetic Algorithm
//
// author: J.I.v.Hemert
// last updated : 07-11-1997
//
// This file implements the class GGAC.
//


#include "gga.h"


extern MLCG rnd;			// Used for random numbers
extern IniFileC inifile;	// Contains all adjustable parameters


void GGAC::Initialize (int numberofobjects, int maxevaluations, int seed)
// Set all genetic algorithm specific parameters and
// initialize the population.
{
	TimerC timer;

	// This random-stuff is used in the mutation, inversion, crossover
	// and the 2-tournament.
	if (inifile.ReadBool ("randomize"))
		randomseed = timer.GetTime ();
	else
		randomseed = seed;

	rnd.seed1 (5);
	rnd.seed2 (randomseed);

	nrofobjects = numberofobjects;
	maxevals = maxevaluations;

	gaparams.PopulationSize = inifile.ReadInt ("populationsize");
	gaparams.N_Crossover = inifile.ReadInt ("crossover");
	gaparams.N_Mutation = inifile.ReadInt ("mutation");
	gaparams.N_Inversion = inifile.ReadInt ("inversion");
	gaparams.AllelMutationProb = inifile.ReadDouble ("allelemutationprob");

	if (gaparams.PopulationSize < gaparams.N_Crossover * 2)
	{
		cerr << "Error: the population size (" << gaparams.PopulationSize << ") is smaller than twice the amount of crossovers (" << gaparams.N_Crossover << ") per generation" << endl;
		exit(2);
	}
	if (gaparams.PopulationSize < gaparams.N_Mutation)
	{
		cerr << "Error: the population size (" << gaparams.PopulationSize << ") is smaller than the amount of mutations (" << gaparams.N_Mutation << ") per generation" << endl;
		exit(2);
	}
	if (gaparams.PopulationSize < gaparams.N_Inversion)
	{
		cerr << "Error: the population size (" << gaparams.PopulationSize << ") is smaller than the amount of inversions (" << gaparams.N_Inversion << ") per generation" << endl;
		exit(2);
	}
	if ((gaparams.AllelMutationProb < 0) || (gaparams.AllelMutationProb > 1))
	{
		cerr << "Error: the parameter AllelMutationProb (" << gaparams.AllelMutationProb << ") should lie between 0 and 1" << endl;
		exit(2);

	}

	debug = inifile.ReadBool ("ggadebug");
	plotdata = inifile.ReadBool ("plotdata");
	printsolutions = inifile.ReadBool ("printsolutions");

	if (printsolutions)
	{
		solutionsfile.open (inifile.ReadString("solutionsfile"));
		if (!solutionsfile.is_open ())
		{
			printsolutions = false;
			cerr << "Warning: could not open " 
					<< inifile.ReadString("solutionsfile") << endl;
		}
	}

	if (plotdata)
	{
		datafile.open (inifile.ReadString("datafile"));
		if (!datafile.is_open ())
		{
			plotdata = false;
			cerr << "Warning: could not open " 
					<< inifile.ReadString("datafile") << endl;
		}
	}

} // Initialize ()


void GGAC::InitializePopulation ()
{
	randomseed++;
	rnd.seed2 (randomseed);
	ColoringT coloringalgorithm = undefined;

	if (strcmp (inifile.ReadString ("coloringalgorithm"), "firstfit") == 0)
			coloringalgorithm = firstfit;
	else
		if (strcmp (inifile.ReadString ("coloringalgorithm"), "smallfirst") == 0)
				coloringalgorithm = smallfirst;
		else
			if (strcmp (inifile.ReadString ("coloringalgorithm"), "largefirst") == 0)
					coloringalgorithm = largefirst;

	population.Initialize (gaparams, nrofobjects, inifile.ReadBool ("populationdebug"), inifile.ReadInt ("k-coloring"), coloringalgorithm);

} // InintializePupulation ()


bool GGAC::Run ()
// Run the genetic algorithm until it solves the problem or until
// it has reached _maxevals_ evaluations. Return true if a solution
// is found.
{
	int gen;

	population.Evaluate ();
	gen = 0;
	if (debug)
		cerr << "\b\b\b\b\b\b\b\b\b\b" 
				<< setw(5) 
				<< population.GetBestFitness ()
				<< setw(5)
				<< GetColorsUsed ();
	
	while ((population.GetBestFitness () > 0) && (population.GetTotalEvaluations ()) < maxevals)
	{
		population.Reproduce ();
		gen++;
		population.Evaluate ();
		if (debug)
			cerr << "\b\b\b\b\b\b\b\b\b\b" 
					<< setw(5) 
					<< population.GetBestFitness ()
					<< setw(5)
					<< GetColorsUsed ();

		if (plotdata)
			datafile << population.GetTotalEvaluations ()  
					 << " " 
					 << population.GetBestFitness () << endl;
	}

	if (plotdata)
		datafile << endl;

	// Print some stats
	if (debug) 
		cerr << "\b\b\b\b\b" 
			 << "total evals = " 
			 << population.GetTotalEvaluations () 
			 << "    bestfitness = " 
			 << population.GetBestFitness () << endl;

	if (printsolutions)
		population.PrintBest (solutionsfile);

	if (population.GetBestFitness () == 0)
		return (true);
	else
		return (false);
} // Run ()


void GGAC::Close ()
{
	if (printsolutions)
		solutionsfile.close ();
	if (plotdata)
		datafile.close ();
} // Close ()


double GGAC::GetColorsUsed ()
{
	return (population.GetColorsUsed ());

} // GetColorsUsed ()


double GGAC::GetTotalEvaluations ()
// Return the totalevaluations done so far by the
// genetic algorithm, no calculations are done.
{
	return (population.GetTotalEvaluations ());

} // GetTotalEvaluation ()


// eof gga.cc