You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 

308 lines
10 KiB

//==============================================================================
//
// Copyright (c) 2002-
// Authors:
// * Dave Parker <david.parker@comlab.ox.ac.uk> (University of Oxford)
//
//------------------------------------------------------------------------------
//
// This file is part of PRISM.
//
// PRISM is free software; you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation; either version 2 of the License, or
// (at your option) any later version.
//
// PRISM is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with PRISM; if not, write to the Free Software Foundation,
// Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
//
//==============================================================================
package explicit;
import java.util.BitSet;
import java.util.List;
import java.util.Map;
import parser.ast.ExpressionTemporal;
import prism.PrismComponent;
import prism.PrismException;
import prism.PrismNotSupportedException;
/**
* Explicit-state model checker for continuous-time Markov decision processes (CTMDPs).
*/
public class CTMDPModelChecker extends ProbModelChecker
{
/**
* Create a new CTMDPModelChecker, inherit basic state from parent (unless null).
*/
public CTMDPModelChecker(PrismComponent parent) throws PrismException
{
super(parent);
}
// Model checking functions
@Override
protected StateValues checkProbBoundedUntil(Model model, ExpressionTemporal expr, MinMax minMax, BitSet statesOfInterest) throws PrismException
{
double uTime;
BitSet b1, b2;
StateValues probs = null;
ModelCheckerResult res = null;
// get info from bounded until
uTime = expr.getUpperBound().evaluateDouble(constantValues);
if (uTime < 0 || (uTime == 0 && expr.upperBoundIsStrict())) {
String bound = (expr.upperBoundIsStrict() ? "<" : "<=") + uTime;
throw new PrismException("Invalid upper bound " + bound + " in time-bounded until formula");
}
// model check operands first for all states
b1 = checkExpression(model, expr.getOperand1(), null).getBitSet();
b2 = checkExpression(model, expr.getOperand2(), null).getBitSet();
// compute probabilities
// a trivial case: "U<=0"
if (uTime == 0) {
// prob is 1 in b2 states, 0 otherwise
probs = StateValues.createFromBitSetAsDoubles(b2, model);
} else {
res = computeBoundedUntilProbs((CTMDP) model, b1, b2, uTime, minMax.isMin());
probs = StateValues.createFromDoubleArray(res.soln, model);
}
return probs;
}
/**
* Compute bounded until probabilities.
* i.e. compute the min/max probability of reaching a state in {@code target},
* within time t, and while remaining in states in @{code remain}.
* @param ctmdp The CTMDP
* @param remain Remain in these states (optional: null means "all")
* @param target Target states
* @param t Bound
* @param min Min or max probabilities (true=min, false=max)
*/
public ModelCheckerResult computeBoundedUntilProbs(CTMDP ctmdp, BitSet remain, BitSet target, double t, boolean min) throws PrismException
{
return computeBoundedReachProbs(ctmdp, remain, target, t, min, null, null);
}
/**
* Compute bounded probabilistic reachability.
* @param ctmdp The CTMDP
* @param remain Remain in these states (optional: null means "all")
* @param target Target states
* @param t Time bound
* @param min Min or max probabilities for (true=min, false=max)
* @param init Initial solution vector - pass null for default
* @param results Optional array of size b+1 to store (init state) results for each step (null if unused)
*/
public ModelCheckerResult computeBoundedReachProbs(CTMDP ctmdp, BitSet remain, BitSet target, double t, boolean min, double init[], double results[]) throws PrismException
{
// TODO: implement until
MDP mdp;
MDPModelChecker mc;
ModelCheckerResult res;
if (!ctmdp.isLocallyUniform())
throw new PrismException("Can't compute bounded reachability probabilities for non-locally uniform CTMDP");
// TODO: check locally uniform
double epsilon = 1e-3;
double q = ctmdp.getMaxExitRate();
int k = (int) Math.ceil((q * t * q * t) / (2 * epsilon));
double tau = t / k;
mainLog.println("q = " + q + ", k = " + k + ", tau = " + tau);
mdp = ctmdp.buildDiscretisedMDP(tau);
mainLog.println(mdp);
mc = new MDPModelChecker(this);
mc.inheritSettings(this);
res = mc.computeBoundedUntilProbs(mdp, null, target, k, min);
return res;
}
/**
* Compute bounded reachability/until probabilities.
* i.e. compute the min/max probability of reaching a state in {@code target},
* within time t, and while remaining in states in @{code remain}.
* @param ctmdp The CTMDP
* @param remain Remain in these states (optional: null means "all")
* @param target Target states
* @param t: Time bound
* @param min Min or max probabilities (true=min, false=max)
* @param init: Initial solution vector - pass null for default
* @param results: Optional array of size b+1 to store (init state) results for each step (null if unused)
*/
public ModelCheckerResult computeBoundedReachProbsOld(CTMDP ctmdp, BitSet remain, BitSet target, double t, boolean min, double init[], double results[]) throws PrismException
{
// TODO: implement until
ModelCheckerResult res = null;
int i, n, iters;
double soln[], soln2[], tmpsoln[], sum[];
long timer;
// Fox-Glynn stuff
FoxGlynn fg;
int left, right;
double q, qt, weights[], totalWeight;
// Start bounded probabilistic reachability
timer = System.currentTimeMillis();
mainLog.println("Starting time-bounded probabilistic reachability...");
// Store num states
n = ctmdp.getNumStates();
// Get uniformisation rate; do Fox-Glynn
q = 99;//ctmdp.unif;
qt = q * t;
mainLog.println("\nUniformisation: q.t = " + q + " x " + t + " = " + qt);
fg = new FoxGlynn(qt, 1e-300, 1e+300, termCritParam / 8.0);
left = fg.getLeftTruncationPoint();
right = fg.getRightTruncationPoint();
if (right < 0) {
throw new PrismException("Overflow in Fox-Glynn computation (time bound too big?)");
}
weights = fg.getWeights();
totalWeight = fg.getTotalWeight();
for (i = left; i <= right; i++) {
weights[i - left] /= totalWeight;
}
mainLog.println("Fox-Glynn: left = " + left + ", right = " + right);
// Create solution vector(s)
soln = new double[n];
soln2 = (init == null) ? new double[n] : init;
sum = new double[n];
// Initialise solution vectors. Use passed in initial vector, if present
if (init != null) {
for (i = 0; i < n; i++)
soln[i] = soln2[i] = target.get(i) ? 1.0 : init[i];
} else {
for (i = 0; i < n; i++)
soln[i] = soln2[i] = target.get(i) ? 1.0 : 0.0;
}
for (i = 0; i < n; i++)
sum[i] = 0.0;
// If necessary, do 0th element of summation (doesn't require any matrix powers)
if (left == 0)
for (i = 0; i < n; i++)
sum[i] += weights[0] * soln[i];
// Start iterations
iters = 1;
while (iters <= right) {
// Matrix-vector multiply and min/max ops
ctmdp.mvMultMinMax(soln, min, soln2, target, true, null);
// Since is globally uniform, can do this? and more?
for (i = 0; i < n; i++)
soln2[i] /= q;
// Store intermediate results if required
// TODO?
// Swap vectors for next iter
tmpsoln = soln;
soln = soln2;
soln2 = tmpsoln;
// Add to sum
if (iters >= left) {
for (i = 0; i < n; i++)
sum[i] += weights[iters - left] * soln[i];
}
iters++;
}
// Print vector (for debugging)
mainLog.println(sum);
// Finished bounded probabilistic reachability
timer = System.currentTimeMillis() - timer;
mainLog.print("Time-bounded probabilistic reachability (" + (min ? "min" : "max") + ")");
mainLog.println(" took " + iters + " iters and " + timer / 1000.0 + " seconds.");
// Return results
res = new ModelCheckerResult();
res.soln = sum;
res.lastSoln = soln2;
res.numIters = iters;
res.timeTaken = timer / 1000.0;
return res;
}
/**
* Compute reachability probabilities.
* i.e. compute the min/max probability of reaching a state in {@code target}.
* @param ctmdp The CTMDP
* @param target Target states
* @param min Min or max probabilities (true=min, false=max)
*/
public ModelCheckerResult computeReachProbs(CTMDP ctmdp, BitSet target, boolean min) throws PrismException
{
throw new PrismNotSupportedException("Not implemented yet");
}
/**
* Construct strategy information for min/max reachability probabilities.
* (More precisely, list of indices of choices resulting in min/max.)
* (Note: indices are guaranteed to be sorted in ascending order.)
* @param ctmdp The CTMDP
* @param state The state to generate strategy info for
* @param target The set of target states to reach
* @param min Min or max probabilities (true=min, false=max)
* @param lastSoln Vector of values from which to recompute in one iteration
*/
public List<Integer> probReachStrategy(CTMDP ctmdp, int state, BitSet target, boolean min, double lastSoln[]) throws PrismException
{
throw new PrismNotSupportedException("Not implemented yet");
}
/**
* Simple test program.
*/
public static void main(String args[])
{
CTMDPModelChecker mc;
CTMDPSimple ctmdp;
ModelCheckerResult res;
BitSet target;
Map<String, BitSet> labels;
boolean min = true;
try {
mc = new CTMDPModelChecker(null);
ctmdp = new CTMDPSimple();
ctmdp.buildFromPrismExplicit(args[0]);
ctmdp.addInitialState(0);
System.out.println(ctmdp);
labels = StateModelChecker.loadLabelsFile(args[1]);
System.out.println(labels);
target = labels.get(args[2]);
if (target == null)
throw new PrismException("Unknown label \"" + args[2] + "\"");
for (int i = 4; i < args.length; i++) {
if (args[i].equals("-min"))
min = true;
else if (args[i].equals("-max"))
min = false;
else if (args[i].equals("-nopre"))
mc.setPrecomp(false);
}
res = mc.computeBoundedReachProbs(ctmdp, null, target, Double.parseDouble(args[3]), min, null, null);
System.out.println(res.soln[0]);
} catch (PrismException e) {
System.out.println(e);
}
}
}