Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance. Project price only 1 $
You can buy this project and download/modify it how often you want.
/*
* Software License Agreement (BSD License)
*
* Copyright 2013 Marc Pujol .
*
* Redistribution and use of this software in source and binary forms, with or
* without modification, are permitted provided that the following conditions
* are met:
*
* Redistributions of source code must retain the above
* copyright notice, this list of conditions and the
* following disclaimer.
*
* Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the
* following disclaimer in the documentation and/or other
* materials provided with the distribution.
*
* Neither the name of IIIA-CSIC, Artificial Intelligence Research Institute
* nor the names of its contributors may be used to
* endorse or promote products derived from this
* software without specific prior written permission of
* IIIA-CSIC, Artificial Intelligence Research Institute
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
package es.csic.iiia.bms.factors;
import es.csic.iiia.bms.MaxOperator;
import es.csic.iiia.bms.util.BestKValuesTracker;
import java.util.Map;
/**
* Factor that selects the best k active neighbor but doesn't constraint others.
*
* This is very similar to the {@link IndependentFactor}, but in this case the utility obtained by
* the factor is that of only the best k active neighbors (instead of all of them).
* Formally, this factor is defined over a sequence of neighbors X = {x_1, ..., x_n}, a
* sequence of corresponding weights B = {b_1, ..., b_n}, b_i > 0, and a maximum number
* of contributing neighbors k. Then, the factor is defined as
*
* where n_a = \sum x_i is the number of active variables.
*
* Messages from this factor can be computed in O(n log(k)) time. First, the maximization
* cases are divided in two groups: those where n_a <= k and those where n_a > k.
*
* @param Type of the factor's identity.
* @author Marc Pujol
*/
public class SaturationKFactor extends IndependentFactor {
private final int k;
/** Tracks the maximum gain from the max(b_i) part */
private BestKValuesTracker max_b;
/** Tracks the maximum gain from the max(b_i + v_i) part (k elements) */
private BestKValuesTracker max_bv_0;
/** Tracks the maximum gain from the max(b_i + v_i) part (k-1 elements) */
private BestKValuesTracker max_bv_1;
/** Tracks the sum of positive messages */
private double sum;
/**
* Build a new saturation factor, that gains utility from at most k active neighbors.
*
* @param k maximum number of neighbors that can contribute to this factor's utility (the
* saturation threshold)
*/
public SaturationKFactor(int k) {
this.k = k;
}
@Override
public void setMaxOperator(MaxOperator maxOperator) {
super.setMaxOperator(maxOperator);
max_b = new BestKValuesTracker(maxOperator, k);
max_bv_0 = new BestKValuesTracker(maxOperator, k-1);
max_bv_1 = new BestKValuesTracker(maxOperator, k-2);
}
/**
* Maximum number of neighbors to select.
*
* @return maximum number of neighbors to select (the k value)
*/
public int getK() {
return k;
}
@Override
protected double eval(Map values) {
BestKValuesTracker chosen = new BestKValuesTracker(getMaxOperator(), k);
// Pick all active neighbors (the TreeSet will evict the ones with lower utilities)
for (T neighbor : getNeighbors()) {
if (values.get(neighbor)) {
chosen.track(neighbor, getPotential(neighbor));
}
}
return chosen.sum();
}
// TODO: Track the number of constraint checks
@Override
public long run() {
final MaxOperator max = getMaxOperator();
max_b.reset();
max_bv_0.reset();
max_bv_1.reset();
// Compute the maximum lists
for (T neighbor : getNeighbors()) {
final double b_i = getPotential(neighbor);
final double v_i = getMessage(neighbor);
final double v_i_negative = max.compare(v_i, 0) >= 0 ? 0 : v_i;
max_b.track(neighbor, b_i + v_i_negative);
final double bv_i = b_i + v_i;
if (max.compare(bv_i, 0) >= 0) {
max_bv_0.track(neighbor, bv_i);
max_bv_1.track(neighbor, bv_i);
}
sum += max.max(v_i, 0);
}
// Compute messages
for (T neighbor : getNeighbors()) {
final double b_i = getPotential(neighbor);
final double v_i = getMessage(neighbor);
final double v_i_positive = max.max(v_i, 0);
final double sum_i = sum - v_i_positive;
final double v_i_negative = max.compare(v_i, 0) >= 0 ? 0 : v_i;
final double m_0_lower = max_bv_0.sumComplementaries(neighbor, b_i + v_i, null);
final double m_0_upper = max_b.sumComplementaries(neighbor, b_i + v_i_negative, null) + sum_i;
final double m_0 = max.max(m_0_lower, m_0_upper);
final double m_1_lower = max_bv_1.sumComplementaries(neighbor, b_i + v_i, null) + b_i;
final double m_1_upper = max_b.sumComplementaries(neighbor, b_i + v_i_negative, b_i) + sum_i;
final double m_1 = max.max(m_1_lower, m_1_upper);
send(m_1 - m_0, neighbor);
}
return 0;
}
}
