Java tutorial
/* LICENSE Copyright (c) 2013-2016, Jesse Hostetler (jessehostetler@gmail.com) All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 2. 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. 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 edu.oregonstate.eecs.mcplan.domains.taxi; import java.io.File; import java.util.ArrayList; import java.util.Arrays; import org.apache.commons.math3.random.MersenneTwister; import org.apache.commons.math3.random.RandomGenerator; import weka.core.Attribute; import weka.core.DenseInstance; import weka.core.Instances; import edu.oregonstate.eecs.mcplan.ActionSpace; import edu.oregonstate.eecs.mcplan.MarkovDecisionProblem; import edu.oregonstate.eecs.mcplan.Pair; import edu.oregonstate.eecs.mcplan.Policy; import edu.oregonstate.eecs.mcplan.abstraction.WekaUtil; import edu.oregonstate.eecs.mcplan.dp.SparseValueIterationSolver; import edu.oregonstate.eecs.mcplan.util.Fn; import edu.oregonstate.eecs.mcplan.util.Generator; /** * @author jhostetler * */ public class TaxiMDP extends MarkovDecisionProblem<TaxiState, TaxiAction> { private final TaxiStateSpace S_; private final TaxiActionSpace A_; public TaxiMDP(final TaxiState s) { S_ = new TaxiStateSpace(s); A_ = new TaxiActionSpace(); } @Override public TaxiStateSpace S() { return S_; } @Override public ActionSpace<TaxiState, TaxiAction> A() { return A_; } private final int[][] neighborhood = new int[][] { { 1, 0 }, { -1, 0 }, { 0, 1 }, { 0, -1 } }; private Pair<ArrayList<int[][]>, ArrayList<Double>> nextStep(final TaxiState current, final int taxi_idx) { final ArrayList<int[][]> succ = new ArrayList<int[][]>(); final ArrayList<Double> prob = new ArrayList<Double>(); final double uniform = 1.0 / (neighborhood.length + 1); final int[] taxi_pos = current.other_taxis[taxi_idx]; final int[] new_pos = Fn.copy(taxi_pos); double stay_prob = uniform; for (final int[] neighbor : neighborhood) { Fn.vplus_inplace(new_pos, neighbor); if (current.isLegalMove(taxi_idx, taxi_pos, new_pos)) { final int[][] new_other = Fn.copy(current.other_taxis); Fn.memcpy(new_other[taxi_idx], new_pos); succ.add(new_other); prob.add(uniform); } else { // If move is illegal, taxi stays put stay_prob += uniform; } Fn.vminus_inplace(new_pos, neighbor); } // Case where taxi doesn't move succ.add(Fn.copy(current.other_taxis)); prob.add(stay_prob); return Pair.makePair(succ, prob); } private Pair<ArrayList<TaxiState>, ArrayList<Double>> otherMoves(final TaxiState s) { ArrayList<TaxiState> succ = new ArrayList<TaxiState>(); ArrayList<Double> prob = new ArrayList<Double>(); succ.add(s); prob.add(1.0); ArrayList<TaxiState> others = new ArrayList<TaxiState>(); ArrayList<Double> other_probs = new ArrayList<Double>(); for (int i = 0; i < s.Nother_taxis; ++i) { for (int si = 0; si < succ.size(); ++si) { // For each successor, compute possible moves for taxi i final TaxiState succ_i = succ.get(si); final Pair<ArrayList<int[][]>, ArrayList<Double>> p = nextStep(succ_i, i); // For each possible move of taxi i, create a new state for (int j = 0; j < p.first.size(); ++j) { final TaxiState sprime = new TaxiState(succ_i); Fn.memcpy(sprime.other_taxis, p.first.get(j)); others.add(sprime); other_probs.add(prob.get(si) * p.second.get(j)); } } // Replace successors with successors including taxi i succ = others; prob = other_probs; others = new ArrayList<TaxiState>(); other_probs = new ArrayList<Double>(); } return Pair.makePair(succ, prob); } @Override public Pair<ArrayList<TaxiState>, ArrayList<Double>> sparseP(final TaxiState s, final TaxiAction a) { final TaxiState sprime = new TaxiState(s); if (a instanceof MoveAction) { final MoveAction move = (MoveAction) a; final int[] new_taxi = Fn.copy(sprime.taxi); new_taxi[0] += move.dx; new_taxi[1] += move.dy; if (s.isLegalMove(s.taxi, new_taxi)) { Fn.memcpy(sprime.taxi, new_taxi); } } else if (a instanceof PickupAction) { if (s.passenger != TaxiState.IN_TAXI && Arrays.equals(s.locations.get(s.passenger), s.taxi)) { sprime.passenger = TaxiState.IN_TAXI; } } else if (a instanceof PutdownAction) { if (s.passenger == TaxiState.IN_TAXI && Arrays.equals(s.locations.get(s.destination), s.taxi)) { sprime.passenger = s.destination; } } else { throw new IllegalArgumentException(a.toString()); } return otherMoves(sprime); } @Override public double[] P(final TaxiState s, final TaxiAction a) { throw new UnsupportedOperationException("Use sparseP()"); } @Override public double P(final TaxiState s, final TaxiAction a, final TaxiState sprime) { throw new UnsupportedOperationException("Use sparseP()"); } @Override public double R(final TaxiState s) { return 0.0; } @Override public double R(final TaxiState s, final TaxiAction a) { if (a instanceof MoveAction) { return -1; } else if (a instanceof PickupAction) { if (s.passenger != TaxiState.IN_TAXI) { if (Arrays.equals(s.locations.get(s.passenger), s.taxi)) { return -1; } } return -10 - 1; } else if (a instanceof PutdownAction) { if (s.passenger == TaxiState.IN_TAXI) { if (Arrays.equals(s.locations.get(s.destination), s.taxi)) { return 20 - 1; } } return -10 - 1; } else { throw new IllegalArgumentException(a.toString()); } } // ----------------------------------------------------------------------- public static void main(final String[] argv) { final int Nother_taxis = 2; final double slip = 0.1; final double discount = 0.9; final RandomGenerator rng = new MersenneTwister(42); final TaxiState template = TaxiWorlds.dietterich2000(rng, Nother_taxis, slip); final TaxiMDP mdp = new TaxiMDP(template); final int Nfeatures = new PrimitiveTaxiRepresentation(template).phi().length; final SparseValueIterationSolver<TaxiState, TaxiAction> vi = new SparseValueIterationSolver<TaxiState, TaxiAction>( mdp, discount); vi.run(); final PrimitiveTaxiRepresenter repr = new PrimitiveTaxiRepresenter(template); final ArrayList<Attribute> attr = new ArrayList<Attribute>(); attr.addAll(repr.attributes()); attr.add(WekaUtil.createNominalAttribute("__label__", mdp.A().cardinality())); final Instances instances = WekaUtil.createEmptyInstances("taxi_" + Nother_taxis + "_pistar", attr); final Policy<TaxiState, TaxiAction> pistar = vi.pistar(); final Generator<TaxiState> g = mdp.S().generator(); while (g.hasNext()) { final TaxiState s = g.next(); pistar.setState(s, 0L); final TaxiAction astar = pistar.getAction(); final double[] phi = new double[Nfeatures + 1]; Fn.memcpy_as_double(phi, new PrimitiveTaxiRepresentation(s).phi(), Nfeatures); phi[Nfeatures] = mdp.A().index(astar); WekaUtil.addInstance(instances, new DenseInstance(1.0, phi)); } WekaUtil.writeDataset(new File("."), instances); } }