Population-based incremental learning

In computer science and machine learning, population-based incremental learning (PBIL) is one of the optimization algorithm, and one of the estimation of distribution algorithm. This is a type of genetic algorithm where the genotype of an entire population is evolved rather than individual members. The algorithm is proposed by Shumeet Baluja in 1994.

Algorithm

In PBIL, genes are represented as real values in the range [0,1], indicating the probability that any particular allele appears in that gene.

The PBIL algorithm is as follows:

1. A population is generated.
2. The fitness of each member is evaluated and ranked.
3. Update population genotype based on fittest individual.
4. Repeat steps 2-3

Source code

This is a part of source code implemented in Java. In the paper, learnRate = 0.1, negLearnRate = 0.075, mutProb = 0.02, and mutShift = 0.05 is used. N = 100 and ITER_COUNT = 1000 is enough for a small problem.

public void optimize() {
    final int totalBits = getTotalBits(domains);
    final double[] probVec = new double[totalBits];
    Arrays.fill(probVec, 0.5);
    bestCost = POSITIVE_INFINITY;
 
    for (int i = 0; i < ITER_COUNT; i++) {
        // Creates N genotypes
        final boolean[][] genoTypes = new boolean[N][totalBits];
        for (boolean[] genoType : genoTypes) {
            for (int k = 0; k < genoType.length; k++) {
                if (rand.nextDouble() < probVec[k])
                    genoType[k] = true;
            }
        }

        // Calculate costs
        final double[] costs = new double[N];
        for (int j = 0; j < N; j++) {
            costs[j] = costFunc.cost(toRealVec(genoTypes[j], domains));
        }

        // Find min and max cost genotypes
        boolean[] minGenoType = null, maxGenoType = null;
        double minCost = POSITIVE_INFINITY, maxCost = NEGATIVE_INFINITY;
        for (int j = 0; j < N; j++) {
            double cost = costs[j];
            if (minCost > cost) {
                minCost = cost;
                minGenoType = genoTypes[j];
            }
            if (maxCost < cost) {
                maxCost = cost;
                maxGenoType = genoTypes[j];
            }
        }

        // Compare with the best cost genotypes
        if (bestCost > minCost) {
            bestCost = minCost;
            bestGenoType = minGenoType;
        }

        // Update the probability vector with max and min cost genotypes
        for (int j = 0; j < totalBits; j++) {
            if (minGenoType[j] == maxGenoType[j]) {
                probVec[j] = probVec[j] * (1d - learnRate) +
                        (minGenoType[j] ? 1d : 0d) * learnRate;
            } else {
                final double learnRate2 = learnRate + negLearnRate;
                probVec[j] = probVec[j] * (1d - learnRate2) +
                        (minGenoType[j] ? 1d : 0d) * learnRate2;
            }
        }

        // Mutation
        for (int j = 0; j < totalBits; j++) {
            if (rand.nextDouble() < mutProb) {
                probVec[j] = probVec[j] * (1d - mutShift) +
                        (rand.nextBoolean() ? 1d : 0d) * mutShift;
            }
        }
    }
}

See also

• Estimation of Distribution Algorithm (EDA)