Редукция в генетическом алгоритме

CDiophantine::CDiophantine(int a, int b, int c, int d, int res) : ca(a), cb(b), cc(c), cd(d), result(res) {}
 
int CDiophantine::Solve() {
    int fitness = -1;
 
    // Generate initial population.
    srand((unsigned)time(NULL));
 
    for (int i = 0; i<MAXPOP; i++) {// Fill the population with numbers between
        for (int j = 0; j<4; j++) {// 0 and the result.
            population[i].alleles[j] = rand() % (result + 1);
        }
    }
 
    if (fitness = CreateFitnesses()) {
        return fitness;
    }
 
    int iterations = 0;// Keep record of the iterations.
    while (fitness != 0 || iterations < 50) {// Repeat until solution found, or over 50 iterations.
        GenerateLikelihoods();// Create the likelihoods.
        CreateNewPopulation();
        if (fitness = CreateFitnesses()) {
            return fitness;
        }
 
        iterations++;
    }
 
    return -1;
}
 
int CDiophantine::Fitness(gene &gn) {
    int total = ca * gn.alleles[0] + cb * gn.alleles[1] + cc * gn.alleles[2] + cd * gn.alleles[3];
 
    return gn.fitness = abs(total - result);
}
 
int CDiophantine::CreateFitnesses() {
    float avgfit = 0;
    int fitness = 0;
    for (int i = 0; i<MAXPOP; i++) {
        fitness = Fitness(population[i]);
        avgfit += fitness;
        if (fitness == 0) {
            return i;
        }
    }
 
    return 0;
}
 
float CDiophantine::MultInv() {
    float sum = 0;
 
    for (int i = 0; i<MAXPOP; i++) {
        sum += 1 / ((float)population[i].fitness);
    }
 
    return sum;
}
 
void CDiophantine::GenerateLikelihoods() {
    float multinv = MultInv();
 
    float last = 0;
    for (int i = 0; i<MAXPOP; i++) {
        population[i].likelihood = last = last + ((1 / ((float)population[i].fitness) / multinv) * 100);
    }
}
 
int CDiophantine::GetIndex(float val) {
    float last = 0;
    for (int i = 0; i<MAXPOP; i++) {
        if (last <= val && val <= population[i].likelihood) return i;
        else last = population[i].likelihood;
    }
 
    return 4;
}
 
gene CDiophantine::Breed(int p1, int p2) {
    int crossover = rand() % 3 + 1;// Create the crossover point (not first).
    int first = rand() % 100;// Which parent comes first?
 
    gene child = population[p1];// Child is all first parent initially.
 
    int initial = 0, final = 3;// The crossover boundaries.
    if (first < 50) initial = crossover;    // If first parent first. start from crossover.
    else final = crossover + 1;// Else end at crossover.
 
    for (int i = initial; i<final; i++) {// Crossover!
        child.alleles[i] = population[p2].alleles[i];
        if (rand() % 101 < 5) child.alleles[i] = rand() % (result + 1);
    }
 
    return child;// Return the kid...
}
 
void CDiophantine::CreateNewPopulation() {
    gene temppop[MAXPOP];
 
    for (int i = 0; i<25; i++) {
        int parent1 = 0, parent2 = 0, iterations = 0;
        while (parent1 == parent2 || population[parent1] == population[parent2]) {
            parent1 = GetIndex((float)(rand() % 101));
            parent2 = GetIndex((float)(rand() % 101));
            if (++iterations > 25) break;
        }
        temppop[i] = Breed(parent1, parent2);// Create a child.
    }
 
    for (int i = 0; i < 25; i++) population[i] = temppop[i];
}