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.
 
 
 
 
 
 

419 lines
14 KiB

//==============================================================================
//
// Copyright (c) 2002-
// Authors:
// * Dave Parker <david.parker@comlab.ox.ac.uk> (University of Oxford, formerly University of Birmingham)
//
//------------------------------------------------------------------------------
//
// 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
//
//==============================================================================
// includes
#include "PrismSparse.h"
#include <math.h>
#include <util.h>
#include <cudd.h>
#include <dd.h>
#include <odd.h>
#include <dv.h>
#include "sparse.h"
#include "prism.h"
#include "PrismNativeGlob.h"
#include "PrismSparseGlob.h"
#include "jnipointer.h"
#include <new>
//------------------------------------------------------------------------------
JNIEXPORT jlong __jlongpointer JNICALL Java_sparse_PrismSparse_PS_1NondetReachReward
(
JNIEnv *env,
jclass cls,
jlong __jlongpointer t, // trans matrix
jlong __jlongpointer ta, // trans action labels
jobject synchs,
jlong __jlongpointer sr, // state rewards
jlong __jlongpointer trr, // transition rewards
jlong __jlongpointer od, // odd
jlong __jlongpointer rv, // row vars
jint num_rvars,
jlong __jlongpointer cv, // col vars
jint num_cvars,
jlong __jlongpointer ndv, // nondet vars
jint num_ndvars,
jlong __jlongpointer g, // 'goal' states
jlong __jlongpointer in, // 'inf' states
jlong __jlongpointer m, // 'maybe' states
jboolean min // min or max probabilities (true = min, false = max)
)
{
// cast function parameters
DdNode *trans = jlong_to_DdNode(t); // trans matrix
DdNode *trans_actions = jlong_to_DdNode(ta); // trans action labels
DdNode *state_rewards = jlong_to_DdNode(sr); // state rewards
DdNode *trans_rewards = jlong_to_DdNode(trr); // transition rewards
ODDNode *odd = jlong_to_ODDNode(od); // reachable states
DdNode **rvars = jlong_to_DdNode_array(rv); // row vars
DdNode **cvars = jlong_to_DdNode_array(cv); // col vars
DdNode **ndvars = jlong_to_DdNode_array(ndv); // nondet vars
DdNode *goal = jlong_to_DdNode(g); // 'goal' states
DdNode *inf = jlong_to_DdNode(in); // 'inf' states
DdNode *maybe = jlong_to_DdNode(m); // 'maybe' states
// mtbdds
DdNode *a, *tmp = NULL;
// model stats
int n, nc, nc_r;
long nnz, nnz_r;
// sparse matrix
NDSparseMatrix *ndsm = NULL, *ndsm_r = NULL;
// vectors
double *sr_vec = NULL, *soln = NULL, *soln2 = NULL, *tmpsoln = NULL, *inf_vec = NULL;
// timing stuff
long start1, start2, start3, stop;
double time_taken, time_for_setup, time_for_iters;
// adversary stuff
int export_adv_enabled = export_adv;
bool adv_loop = false;
FILE *fp_adv = NULL;
int adv_j;
bool adv_new;
int *adv = NULL;
// action info
int *actions;
jstring *action_names_jstrings;
const char** action_names;
int num_actions;
// misc
int i, j, k, k_r, l1, h1, l2, h2, l2_r, h2_r, iters;
double d1, d2, x, sup_norm, kb, kbt;
bool done, first;
// exception handling around whole function
try {
// start clocks
start1 = start2 = util_cpu_time();
// get number of states
n = odd->eoff + odd->toff;
// filter out rows (goal states and infinity states) from matrix
Cudd_Ref(trans);
Cudd_Ref(maybe);
a = DD_Apply(ddman, APPLY_TIMES, trans, maybe);
// also remove goal and infinity states from state rewards vector
Cudd_Ref(state_rewards);
Cudd_Ref(maybe);
state_rewards = DD_Apply(ddman, APPLY_TIMES, state_rewards, maybe);
// and from transition rewards matrix
Cudd_Ref(trans_rewards);
Cudd_Ref(maybe);
trans_rewards = DD_Apply(ddman, APPLY_TIMES, trans_rewards, maybe);
// build sparse matrix (probs)
PS_PrintToMainLog(env, "\nBuilding sparse matrix (transitions)... ");
ndsm = build_nd_sparse_matrix(ddman, a, rvars, cvars, num_rvars, ndvars, num_ndvars, odd);
// get number of transitions/choices
nnz = ndsm->nnz;
nc = ndsm->nc;
kb = (nnz*12.0+nc*4.0+n*4.0)/1024.0;
kbt = kb;
// print out info
PS_PrintToMainLog(env, "[n=%d, nc=%d, nnz=%d, k=%d] ", n, nc, nnz, ndsm->k);
PS_PrintMemoryToMainLog(env, "[", kb, "]\n");
// if needed, and if info is available, build a vector of action indices for the MDP
actions = NULL;
if (export_adv_enabled != EXPORT_ADV_NONE) {
if (trans_actions != NULL) {
PS_PrintToMainLog(env, "Building action information... ");
// first need to filter out unwanted rows
Cudd_Ref(trans_actions);
Cudd_Ref(maybe);
tmp = DD_Apply(ddman, APPLY_TIMES, trans_actions, maybe);
// then convert to a vector of integer indices
actions = build_nd_action_vector(ddman, a, tmp, ndsm, rvars, cvars, num_rvars, ndvars, num_ndvars, odd);
Cudd_RecursiveDeref(ddman, tmp);
kb = n*4.0/1024.0;
kbt += kb;
PS_PrintMemoryToMainLog(env, "[", kb, "]\n");
// also extract list of action names from 'synchs'
get_string_array_from_java(env, synchs, action_names_jstrings, action_names, num_actions);
} else {
PS_PrintWarningToMainLog(env, "Action labels are not available for adversary generation.", export_adv_filename);
}
}
// build sparse matrix (rewards)
PS_PrintToMainLog(env, "Building sparse matrix (transition rewards)... ");
ndsm_r = build_sub_nd_sparse_matrix(ddman, a, trans_rewards, rvars, cvars, num_rvars, ndvars, num_ndvars, odd);
// get number of transitions/choices
nnz_r = ndsm_r->nnz;
nc_r = ndsm_r->nc;
// print out info
PS_PrintToMainLog(env, "[n=%d, nc=%d, nnz=%d, k=%d] ", n, nc_r, nnz_r, ndsm_r->k);
kb = (nnz_r*12.0+nc_r*4.0+n*4.0)/1024.0;
kbt += kb;
PS_PrintMemoryToMainLog(env, "[", kb, "]\n");
// get vector for state rewards
PS_PrintToMainLog(env, "Creating vector for state rewards... ");
sr_vec = mtbdd_to_double_vector(ddman, state_rewards, rvars, num_rvars, odd);
kb = n*8.0/1024.0;
kbt += kb;
PS_PrintMemoryToMainLog(env, "[", kb, "]\n");
// get vector for yes
PS_PrintToMainLog(env, "Creating vector for inf... ");
inf_vec = mtbdd_to_double_vector(ddman, inf, rvars, num_rvars, odd);
kb = n*8.0/1024.0;
kbt += kb;
PS_PrintMemoryToMainLog(env, "[", kb, "]\n");
// create solution/iteration vectors
PS_PrintToMainLog(env, "Allocating iteration vectors... ");
soln = new double[n];
soln2 = new double[n];
kb = n*8.0/1024.0;
kbt += 2*kb;
PS_PrintMemoryToMainLog(env, "[2 x ", kb, "]\n");
// if required, create storage for adversary and initialise
if (export_adv_enabled != EXPORT_ADV_NONE) {
PS_PrintToMainLog(env, "Allocating adversary vector... ");
adv = new int[n];
kb = n*sizeof(int)/1024.0;
kbt += kb;
PS_PrintMemoryToMainLog(env, "[", kb, "]\n");
// Initialise all entries to -1 ("don't know")
for (i = 0; i < n; i++) {
adv[i] = -1;
}
}
// print total memory usage
PS_PrintMemoryToMainLog(env, "TOTAL: [", kbt, "]\n");
// initial solution is infinity in 'inf' states, zero elsewhere
for (i = 0; i < n; i++) {
soln[i] = (inf_vec[i] > 0) ? HUGE_VAL : 0.0;
}
// get setup time
stop = util_cpu_time();
time_for_setup = (double)(stop - start2)/1000;
start2 = stop;
start3 = stop;
// start iterations
iters = 0;
done = false;
PS_PrintToMainLog(env, "\nStarting iterations...\n");
// open file to store adversary (if required)
if (export_adv_enabled != EXPORT_ADV_NONE) {
fp_adv = fopen(export_adv_filename, "w");
if (fp_adv) {
fprintf(fp_adv, "%d ?\n", n);
} else {
PS_PrintWarningToMainLog(env, "Adversary generation cancelled (could not open file \"%s\").", export_adv_filename);
export_adv_enabled = EXPORT_ADV_NONE;
}
}
// store local copies of stuff
// firstly for transition matrix
double *non_zeros = ndsm->non_zeros;
unsigned char *row_counts = ndsm->row_counts;
int *row_starts = (int *)ndsm->row_counts;
unsigned char *choice_counts = ndsm->choice_counts;
int *choice_starts = (int *)ndsm->choice_counts;
bool use_counts = ndsm->use_counts;
unsigned int *cols = ndsm->cols;
// and then for transition rewards matrix
// (note: we don't need row_counts/row_starts for
// this since choice structure mirrors transition matrix)
double *non_zeros_r = ndsm_r->non_zeros;
unsigned char *choice_counts_r = ndsm_r->choice_counts;
int *choice_starts_r = (int *)ndsm_r->choice_counts;
bool use_counts_r = ndsm_r->use_counts;
unsigned int *cols_r = ndsm_r->cols;
while (!done && iters < max_iters) {
iters++;
// do matrix multiplication and min/max
h1 = h2 = h2_r = 0;
// loop through states
for (i = 0; i < n; i++) {
d1 = 0.0; // initial value doesn't matter
first = true; // (because we also remember 'first')
adv_new = false;
// get pointers to nondeterministic choices for state i
if (!use_counts) { l1 = row_starts[i]; h1 = row_starts[i+1]; }
else { l1 = h1; h1 += row_counts[i]; }
// loop through those choices
for (j = l1; j < h1; j++) {
// compute the reward value for state i for this iteration
// start with state reward for this state
d2 = sr_vec[i];
// get pointers to transitions
if (!use_counts) { l2 = choice_starts[j]; h2 = choice_starts[j+1]; }
else { l2 = h2; h2 += choice_counts[j]; }
// and get pointers to transition rewards
if (!use_counts_r) { l2_r = choice_starts_r[j]; h2_r = choice_starts_r[j+1]; }
else { l2_r = h2_r; h2_r += choice_counts_r[j]; }
// loop through transitions
for (k = l2; k < h2; k++) {
// find corresponding transition reward if any
k_r = l2_r; while (k_r < h2_r && cols_r[k_r] != cols[k]) k_r++;
// if there is one, add reward * prob to reward value
if (k_r < h2_r) { d2 += non_zeros_r[k_r] * non_zeros[k]; k_r++; }
// add prob * corresponding reward from previous iteration
d2 += non_zeros[k] * soln[cols[k]];
}
// see if this value is the min/max so far
if (first || (min&&(d2<d1)) || (!min&&(d2>d1))) {
d1 = d2;
// if adversary generation is enabled, remember optimal choice
if (export_adv_enabled != EXPORT_ADV_NONE) {
// for max, only remember strictly better choices
// (this resolves problems with end components)
if (!min) {
if (adv[i] == -1 || (d1>soln[i])) {
adv[i] = j;
}
}
// for min, this is straightforward
// (in fact, could do it at the end of value iteration, but we don't)
else {
adv[i] = j;
}
}
}
first = false;
}
// set vector element
// (if there were no choices from this state, reward is zero/infinity)
soln2[i] = (h1 > l1) ? d1 : inf_vec[i] > 0 ? HUGE_VAL : 0;
}
// check convergence
sup_norm = 0.0;
for (i = 0; i < n; i++) {
x = fabs(soln2[i] - soln[i]);
if (term_crit == TERM_CRIT_RELATIVE) {
x /= soln2[i];
}
if (x > sup_norm) sup_norm = x;
}
if (sup_norm < term_crit_param) {
done = true;
}
// print occasional status update
if ((util_cpu_time() - start3) > UPDATE_DELAY) {
PS_PrintToMainLog(env, "Iteration %d: max %sdiff=%f", iters, (term_crit == TERM_CRIT_RELATIVE)?"relative ":"", sup_norm);
PS_PrintToMainLog(env, ", %.2f sec so far\n", ((double)(util_cpu_time() - start2)/1000));
start3 = util_cpu_time();
}
// prepare for next iteration
tmpsoln = soln;
soln = soln2;
soln2 = tmpsoln;
// if we're done, but adversary generation is required, go round once more
if (done && adv) adv_loop = !adv_loop;
}
// Traverse matrix to extract adversary
if (export_adv_enabled != EXPORT_ADV_NONE) {
h1 = h2 = 0;
for (i = 0; i < n; i++) {
if (!use_counts) { l1 = row_starts[i]; h1 = row_starts[i+1]; }
else { l1 = h1; h1 += row_counts[i]; }
// Have to loop through all choices (to compute offsets)
for (j = l1; j < h1; j++) {
if (!use_counts) { l2 = choice_starts[j]; h2 = choice_starts[j+1]; }
else { l2 = h2; h2 += choice_counts[j]; }
// But only output a choice if it is in the adversary
if (j == adv[i]) {
for (k = l2; k < h2; k++) {
switch (export_adv_enabled) {
case EXPORT_ADV_DTMC:
fprintf(fp_adv, "%d %d %g", i, cols[k], non_zeros[k]); break;
case EXPORT_ADV_MDP:
fprintf(fp_adv, "%d 0 %d %g", i, cols[k], non_zeros[k]); break;
}
if (actions != NULL) fprintf(fp_adv, " %s", actions[j]>0?action_names[actions[j]-1]:"");
fprintf(fp_adv, "\n");
}
}
}
}
}
// stop clocks
stop = util_cpu_time();
time_for_iters = (double)(stop - start2)/1000;
time_taken = (double)(stop - start1)/1000;
// print iterations/timing info
PS_PrintToMainLog(env, "\nIterative method: %d iterations in %.2f seconds (average %.6f, setup %.2f)\n", iters, time_taken, time_for_iters/iters, time_for_setup);
// if the iterative method didn't terminate, this is an error
if (!done) { delete soln; soln = NULL; PS_SetErrorMessage("Iterative method did not converge within %d iterations.\nConsider using a different numerical method or increasing the maximum number of iterations", iters); }
// close file to store adversary (if required)
if (export_adv_enabled != EXPORT_ADV_NONE) {
fclose(fp_adv);
PS_PrintToMainLog(env, "\nAdversary written to file \"%s\".\n", export_adv_filename);
}
// catch exceptions: register error, free memory
} catch (std::bad_alloc e) {
PS_SetErrorMessage("Out of memory");
if (soln) delete[] soln;
soln = 0;
}
// free memory
if (a) Cudd_RecursiveDeref(ddman, a);
if (state_rewards) Cudd_RecursiveDeref(ddman, state_rewards);
if (trans_rewards) Cudd_RecursiveDeref(ddman, trans_rewards);
if (ndsm) delete ndsm;
if (ndsm_r) delete ndsm_r;
if (inf_vec) delete[] inf_vec;
if (sr_vec) delete[] sr_vec;
if (soln2) delete[] soln2;
if (adv) delete[] adv;
if (actions != NULL) {
delete[] actions;
release_string_array_from_java(env, action_names_jstrings, action_names, num_actions);
}
return ptr_to_jlong(soln);
}
//------------------------------------------------------------------------------