static double best_fit(args_t *args, dist_t *dist, int ngauss, double *params)
{
if ( ngauss==1 )
{
gauss_fit(dist,ngauss,params);
params[1] *= params[1];
return eval_fit(dist->dat.nvals, dist->dat.xvals, dist->dat.yvals, ngauss,params);
}
int i, j, n = 3;
int ipk = 3*(ngauss-1);
double delta = 0.5 * (params[ipk] - params[0]) / n;
double best_params[9], tmp_params[9], best_fit = HUGE_VAL;
for (i=0; i<n; i++)
{
memcpy(tmp_params,params,sizeof(double)*ngauss*3);
tmp_params[0] += delta*i;
tmp_params[ipk] -= delta*i;
if ( gauss_fit(dist,ngauss,tmp_params)<0 ) continue; // did not converge
for (j=0; j<ngauss; j++) tmp_params[j*3+1] *= tmp_params[j*3+1];
// From the nature of the data, we can assume that in presence of
// multiple peaks they will be placed symmetrically around 0.5. Also
// their size should be about the same. We evaluate the fit with this
// in mind.
double dx = fabs(0.5 - tmp_params[0]) + fabs(tmp_params[ipk] - 0.5);
tmp_params[0] = 0.5 - dx*0.5;
tmp_params[ipk] = 0.5 + dx*0.5;
double fit = eval_fit(dist->dat.nvals, dist->dat.xvals, dist->dat.yvals, ngauss, tmp_params);
if ( best_fit < fit ) continue; // worse than previous
best_fit = fit;
memcpy(best_params,tmp_params,sizeof(double)*ngauss*3);
}
memcpy(params,best_params,sizeof(double)*ngauss*3);
return best_fit;
}
/**
* GA with a single fitness function
*/
int evolve_both(int* first, int* second, int* winner,
int* loser, int* population) {
int round = 0;
float fitness_first, fitness_second;
int first_bitflipped[GENOTYPE_LENGTH],
second_bitflipped[GENOTYPE_LENGTH];
do {
round++;
// generate the complementary individuals
bitflip(first, first_bitflipped);
bitflip(second, second_bitflipped);
// mark them as tested in the population
population[hash(first)] = 1;
population[hash(second)] = 1;
// evaluate first and its complementary
fitness_first = eval_fit(first, first_bitflipped);
// evaluate second
fitness_second = eval_fit(second, second_bitflipped);
// optimal solution?
if ( fitness_first == 0.0 )
{
copy_individual(first, winner);
copy_individual(second, loser);
break;
}
if ( fitness_second == 0.0 )
{
copy_individual(second, winner);
copy_individual(first, loser);
break;
}
if ( fitness_first < fitness_second )
{
// first wins
// mutate loser
if ( mutate_1gene(second, population) < 0 )
{
// if no more mutations are possible
copy_individual(first, winner);
copy_individual(second, loser);
break;
}
}
else
{
// second wins
// mutate loser
if ( mutate_1gene(first, population) < 0 )
{
copy_individual(second, winner);
copy_individual(first, loser);
break;
}
}
} while ( 1 );
return round;
}
