void Evnav::makeGraph(Graph &g, EvnavRequest &req, VertexId srcId, VertexId dstId)
{
Trip trip;
// FIXME: move to EvnavRequest?
double SOC_dyn = req.m_SOC_max - req.m_SOC_min;
double batt_act = req.m_battery * req.m_SOC_act; // kWh
double batt_dyn = req.m_battery * SOC_dyn;
// Add edge from the source to all chargers
// FIXME: create waypoint class with chargers that extends it
for (Charger &a : m_provider->chargers()) {
if (computeTrip(req.m_src, a.loc(), trip) == Status::Ok) {
double e = computeEnergy(trip, req.m_efficiency);
if (e < batt_act) {
Edge edge(srcId, a.id(), (double)trip.time_s);
edge.m_travel_time = trip.time_s;
edge.m_dist = trip.dist_m;
edge.m_energy = e;
g.addEdge(edge);
}
}
}
// Add all intermediate chargers
chargerMatrix([&](Charger &a, Charger &b, Trip &t) {
// do not add edges between chargers that are too close
if (t.dist_m < 1000) {
return;
}
double e = computeEnergy(t, req.m_efficiency);
if (e < batt_dyn) {
double total_time = computeTripTimeWithCharging(t, e, req.m_power_avg);
Edge edge(a.id(), b.id(), total_time);
edge.m_energy = e;
edge.m_charge_time = computeChargingTime(e, req.m_power_avg);
edge.m_dist = t.dist_m;
edge.m_travel_time = t.time_s;
g.addEdge(edge);
}
});
// Add edge from all chargers to the destination
for (Charger &a : m_provider->chargers()) {
if (computeTrip(a.loc(), req.m_dst, trip) == Status::Ok) {
double e = computeEnergy(trip, req.m_efficiency);
if (e < batt_dyn) {
double total_time = computeTripTimeWithCharging(trip, e, req.m_power_avg);
Edge edge(a.id(), dstId, total_time);
edge.m_energy = e;
edge.m_charge_time = computeChargingTime(e, req.m_power_avg);
edge.m_dist = trip.dist_m;
edge.m_travel_time = trip.time_s;
g.addEdge(edge);
}
}
}
}
double Model::getEnergy()
{
double energy = 0;
for (unsigned int i=0;i<particles.size();i++)
#ifndef ISOTROPIC
energy += computeEnergy(i, &particles[i].position[0], &particles[i].orientation[0]);
#else
energy += computeEnergy(i, &particles[i].position[0]);
#endif
return energy/(2*particles.size());
}
Real
AuxChem::computeValue()
{
Real matrix_energy(0.0);
Real precip_energy(0.0);
Real differential(0.0);
matrix_energy = computeEnergy(_coupled_cons[_qp], _coupled_noncons[_qp], true);
precip_energy = computeEnergy(_precip_conserved, _precip_nonconserved, false);
differential = computeDifferential(_coupled_cons[_qp], _coupled_noncons[_qp]);
return (matrix_energy - precip_energy + differential);
}
inline real
AngularUniquePotentialTemplate< Derived >::
computeEnergy(const Particle &p1, const Particle &p2, const Particle &p3, real theta0) const {
Real3D r12 = p1.position() - p2.position();
Real3D r32 = p3.position() - p2.position();
return computeEnergy(r12, r32, theta0);
}
inline real
AngularUniquePotentialTemplate< Derived >::
computeEnergy(const Real3D& r12, const Real3D& r32, real theta0) const {
real dist12Sqr = r12.sqr();
real dist32Sqr = r32.sqr();
real cos_theta = r12 * r32 / (sqrt(dist12Sqr) * sqrt(dist32Sqr));
return computeEnergy(acos(cos_theta), theta0);
}
int SimulationIsingModel::attemptFlip(int i)
{
double old_e = computeEnergy(i);
_spin[i] *= -1.0;
double new_e = computeEnergy(i);
//std::cerr << old_e << "\t" << new_e << "\t" << exp(-_beta*(new_e - old_e)) << std::endl;
if (new_e < old_e) {
//std::cerr << 1 << std::endl;
return 1;
}
else if (_rng.randDouble() < exp(-_beta*(new_e - old_e))) {
//std::cerr << 1 << std::endl;
return 1;
}
else {
//std::cerr << 0 << std::endl;
_spin[i] *= -1.0;
return 0;
}
}
void PisdWdF<EvalT, Traits>::
evaluateFields(typename Traits::EvalData workset)
{
ScalarT kappa;
ScalarT mu;
// Leading dimension of 1 added so we can use Intrepid::det
Intrepid::FieldContainer<EnergyFadType> F(1,numDims,numDims);
// Allocate F ( = defgrad of derivative types) and seed with identity derivs
for (std::size_t i=0; i < numDims; ++i)
{
for (std::size_t j=0; j < numDims; ++j)
{
F(0,i,j) = EnergyFadType(numDims*numDims, 0.0); // 0.0 will be overwriten below
F(0,i,j).fastAccessDx(i*numDims + j) = 1.0;
}
}
for (std::size_t cell=0; cell < workset.numCells; ++cell) {
for (std::size_t qp=0; qp < numQPs; ++qp) {
kappa = elasticModulus(cell,qp) / ( 3. * ( 1. - 2. * poissonsRatio(cell,qp) ) );
mu = elasticModulus(cell,qp) / ( 2. * ( 1. + poissonsRatio(cell,qp) ) );
// Fill F with defgrad for value. Derivs already seeded with identity.
for (std::size_t i=0; i < numDims; ++i)
for (std::size_t j=0; j < numDims; ++j)
F(0,i,j).val() = defgrad(cell, qp, i, j);
// Call energy funtional (can make a library of these)
EnergyFadType W = computeEnergy(kappa, mu, F);
// Extract stress from derivs of energy
for (std::size_t i=0; i < numDims; ++i)
for (std::size_t j=0; j < numDims; ++j)
P(cell, qp, i, j) = W.fastAccessDx(i*numDims + j);
}
}
}
double GI_MF::run(labelType* inferredLabels,
int id,
size_t maxiter,
labelType* nodeLabelsGroundTruth,
bool computenergyAtEachIteration,
double* _loss)
{
bool useLossFunction = lossPerLabel!=0;
string paramMSRC;
Config::Instance()->getParameter("msrc", paramMSRC);
bool useMSRC = paramMSRC.c_str()[0] == '1';
bool replaceVoidMSRC = false;
labelType voidLabel = 0;
labelType moutainLabel = 0;
labelType horseLabel = 0;
if(!useLossFunction && useMSRC) {
Config::Instance()->getParameter("msrc_replace_void", paramMSRC);
replaceVoidMSRC = paramMSRC.c_str()[0] == '1';
voidLabel = classIdxToLabel[0];
moutainLabel = classIdxToLabel[4161600];
horseLabel = classIdxToLabel[8323328];
printf("[GI_MF] MSRC void=%d, moutain=%d, horse=%d\n",
(int)voidLabel, (int)moutainLabel, (int)horseLabel);
} else {
printf("[GI_MF] Do not replace void labels\n");
}
int sid = 0;
double maxScore = 0;
double totalScore_old = 0;
double totalScore = 10;
supernode *s;
const map<int, supernode* >& _supernodes = slice->getSupernodes();
ulong nSupernodes = slice->getNbSupernodes();
// check if memory was already allocated for believes
if(!believes) {
believes = new double*[nSupernodes];
for (uint k = 0; k < nSupernodes; ++k) {
believes[k] = new double[param->nClasses];
}
}
double maxPotential;
computeNodePotentials(believes, maxPotential);
double scale = 1.0;
if (fabs(maxPotential) > 1e-30) {
scale = MAX_POTENTIAL/maxPotential;
}
INFERENCE_PRINT("[gi_MF] maxPotential=%g, scale=%g\n", maxPotential, scale);
#if EXP_DOMAIN
INFERENCE_PRINT("[gi_MF] Computing in exp domain\n");
// go to exponential domain
for (uint sid = 0; sid < nSupernodes; ++sid) {
for(int c = 0; c < param->nClasses; ++c) {
believes[sid][c] = std::exp(believes[sid][c]*scale);
}
}
#else
INFERENCE_PRINT("[gi_MF] Computing in log domain\n");
// stay in log domain
for (uint sid = 0; sid < nSupernodes; ++sid) {
for(int c = 0; c < param->nClasses; ++c) {
believes[sid][c] *= scale;
}
}
#endif
//exportBelieves("believes0");
for(map<int, supernode* >::const_iterator its = _supernodes.begin();
its != _supernodes.end(); its++) {
sid = its->first;
inferredLabels[sid] = 0;
}
// random selection
for(uint iter = 0; iter < maxiter && (totalScore - totalScore_old) > 1.0; ++iter) {
printf("[GI_MF] Iteration %d/%ld\n", iter, maxiter);
totalScore_old = totalScore;
totalScore = 0;
// Select a pixel at random
//int sid = gsl_rng_uniform_int(r, nSupernodes);
int nLabelsChanged = 0;
double* bs = 0;
double maxBelief = 0;
for(map<int, supernode* >::const_iterator its = _supernodes.begin();
its != _supernodes.end(); its++) {
sid = its->first;
s = its->second;
bs = believes[sid];
if(param->nClasses != 2) {
for(int c = 0; c < (int)param->nClasses; c++) {
double unaryPotential = computeUnaryPotential(slice, sid, c) * scale;
double pairwiseBelief = 0;
if(param->includeLocalEdges) {
vector < supernode* >* lNeighbors = &(s->neighbors);
for(vector < supernode* >::iterator itN = lNeighbors->begin();
itN != lNeighbors->end(); itN++) {
// set edges once
if(its->first < (*itN)->id) {
continue;
}
#if USE_LONG_RANGE_EDGES
double pairwisePotential = computePairwisePotential_distance(slice, s, (*itN),
c, inferredLabels[(*itN)->id]);
#else
double pairwisePotential = computePairwisePotential(slice, s, (*itN),
c, inferredLabels[(*itN)->id]);
#endif
pairwisePotential *= scale;
pairwiseBelief += believes[(*itN)->id][c] * pairwisePotential;
}
}
#if EXP_DOMAIN
bs[c] = exp(unaryPotential)*exp(pairwiseBelief);
#else
bs[c] = unaryPotential + pairwiseBelief;
#endif
if(maxBelief < bs[c]) {
maxBelief = bs[c];
}
}
} else {
int c = T_FOREGROUND;
#if EXP_DOMAIN
bs[T_FOREGROUND] = 1;
#else
bs[T_FOREGROUND] = 0;
#endif
c = T_BACKGROUND;
double unaryPotential = computeUnaryPotential(slice, sid, c) * scale;
//printf("unaryPotential %d %g\n", sid, unaryPotential);
double pairwiseBelief = 0;
if(param->includeLocalEdges) {
vector < supernode* >* lNeighbors = &(s->neighbors);
for(vector < supernode* >::iterator itN = lNeighbors->begin();
itN != lNeighbors->end(); itN++) {
// set edges once
if(its->first < (*itN)->id) {
continue;
}
#if USE_LONG_RANGE_EDGES
double pairwisePotential = computePairwisePotential_distance(slice, s, (*itN),
c, inferredLabels[(*itN)->id]);
#else
double pairwisePotential = computePairwisePotential(slice, s, (*itN),
c, inferredLabels[(*itN)->id]);
#endif
pairwisePotential *= scale;
pairwiseBelief += believes[(*itN)->id][c] * pairwisePotential;
//printf("pairwisePotential %d %d %d %g %g %g\n", sid, (*itN)->id, c, pairwisePotential, believes[(*itN)->id][c], pairwiseBelief);
}
}
#if EXP_DOMAIN
bs[c] = exp(unaryPotential)*exp(pairwiseBelief);
#else
bs[c] = unaryPotential + pairwiseBelief;
#endif
//printf("bs[%d] %g\n",c,bs[c]);
if(maxBelief < bs[c]) {
maxBelief = bs[c];
}
}
if(useLossFunction) {
// loss function
for(int c = 0; c < (int)param->nClasses; c++) {
if(c != groundTruthLabels[sid]) {
double _loss = lossPerLabel[groundTruthLabels[sid]] * scale;
// add loss of the ground truth label
#if EXP_DOMAIN
bs[c] *= exp(_loss);
#else
bs[c] += _loss;
#endif
if(maxBelief < bs[c]) {
maxBelief = bs[c];
}
}
}
}
// pick max
maxScore = bs[inferredLabels[sid]];
for(int c = 0; c < (int)param->nClasses; c++) {
if(!replaceVoidMSRC || (c != voidLabel && c != moutainLabel && c != horseLabel)) {
if(maxScore < bs[c]) {
maxScore = bs[c];
inferredLabels[sid] = c;
++nLabelsChanged;
}
}
}
totalScore += maxScore;
}
printf("[GI_MF] Iteration %d/%ld. Total score = %g. nLabelsChanged = %d\n", iter, maxiter,
totalScore, nLabelsChanged);
if(maxBelief > MAX_POTENTIAL) {
// normalize believes
printf("[GI_MF] Normalizing believes. maxBelief = %g\n", maxBelief);
for (uint sid = 0; sid < nSupernodes; ++sid) {
for(int c = 0; c < param->nClasses; ++c) {
believes[sid][c] /= maxBelief;
}
}
}
//exportBelieves("believes_last");
}
// cleaning
return computeEnergy(inferredLabels);
}
jdoubleArray Java_edu_cornell_audioProbe_AudioManager_features(JNIEnv* env, jobject javaThis, jshortArray array) {
//void Java_edu_cornell_audioProbe_AudioManager_features(JNIEnv* env, jobject javaThis, jshortArray array) {
(*env)->GetShortArrayRegion(env, array, 0, FRAME_LENGTH, buf);
normalize_data();
// apply window
computeHamming();
// computeFwdFFT
kiss_fftr(cfgFwd, normalizedData, fftx);
//compute power spectrum
computePowerSpec(fftx, powerSpec, FFT_LENGTH);
//compute magnitude spectrum
computeMagnitudeSpec(powerSpec, magnitudeSpec, FFT_LENGTH);
// compute total energy
energy = computeEnergy(powerSpec, FFT_LENGTH) / FFT_LENGTH;
//compute Spectral Entropy
computeSpectralEntropy2(magnitudeSpec, FFT_LENGTH);
//compute auto-correlation peaks
computeAutoCorrelationPeaks2(powerSpec, powerSpecCpx, NOISE_LEVEL, FFT_LENGTH);
//data output
////return data as variable size array caused by variable autocorrelation information.
jdoubleArray featureVector = (*env)->NewDoubleArray(env,6 + 2*numAcorrPeaks + 2 + LOOK_BACK_LENGTH);
//jdoubleArray featureVector = (*env)->NewDoubleArray(env,6 + 2*numAcorrPeaks + 2 + LOOK_BACK_LENGTH + FFT_LENGTH);
featuresValuesTemp[0] = numAcorrPeaks; //autocorrelation values
featuresValuesTemp[1] = maxAcorrPeakVal;
featuresValuesTemp[2] = maxAcorrPeakLag;
featuresValuesTemp[3] = spectral_entropy;
featuresValuesTemp[4] = rel_spectral_entropy;
featuresValuesTemp[5] = energy;
//gaussian distribution
//test the gaussian distribution with some dummy values first
x[0] = maxAcorrPeakVal;
x[1] = numAcorrPeaks;
x[2] = rel_spectral_entropy;
/*
emissionVoiced = computeMvnPdf(x,mean_voiced, inv_cov_voiced, denom_gauss_voiced);
emissionUnvoiced = computeMvnPdf(x,mean_unvoiced, inv_cov_unvoiced, denom_gauss_unvoiced);
*/
inferenceResult = getViterbiInference(x,featureAndInference);
memcpy( featuresValuesTemp+6, featureAndInference, (2+LOOK_BACK_LENGTH)*sizeof(double) ); //observation probabilities, inferences
//put auto correlation values in the string
memcpy( featuresValuesTemp+6+2+LOOK_BACK_LENGTH, acorrPeakValueArray, numAcorrPeaks*sizeof(double) );
memcpy( featuresValuesTemp+6+numAcorrPeaks+2+LOOK_BACK_LENGTH, acorrPeakLagValueArray, numAcorrPeaks*sizeof(double) );
//memcpy( featuresValuesTemp+6+numAcorrPeaks+numAcorrPeaks+2+LOOK_BACK_LENGTH, magnSpect, FFT_LENGTH*sizeof(double) );
(*env)->SetDoubleArrayRegion( env, featureVector, 0, 6 + numAcorrPeaks*2 + 2 + LOOK_BACK_LENGTH, (const jdouble*)featuresValuesTemp );
//(*env)->SetDoubleArrayRegion( env, featureVector, 0, 6 + numAcorrPeaks*2 + 2 + LOOK_BACK_LENGTH + FFT_LENGTH, (const jdouble*)featuresValuesTemp );
return featureVector;
}
int main(int argc, char *argv[])
{
printf("%s Starting...\n\n", argv[0]);
try
{
std::string sFilename;
char *filePath = sdkFindFilePath("person.txt", argv[0]);
if (filePath)
{
sFilename = filePath;
}
else
{
printf("Error %s was unable to find person.txt\n", argv[0]);
exit(EXIT_FAILURE);
}
cudaDeviceInit(argc, (const char **)argv);
printfNPPinfo(argc, argv);
if (g_bQATest == false && (g_nDevice == -1) && argc > 1)
{
sFilename = argv[1];
}
// if we specify the filename at the command line, then we only test sFilename
int file_errors = 0;
std::ifstream infile(sFilename.data(), std::ifstream::in);
if (infile.good())
{
std::cout << "imageSegmentationNPP opened: <" << sFilename.data() << "> successfully!" << std::endl;
file_errors = 0;
infile.close();
}
else
{
std::cout << "imageSegmentationNPP unable to open: <" << sFilename.data() << ">" << std::endl;
file_errors++;
infile.close();
}
if (file_errors > 0)
{
exit(EXIT_FAILURE);
}
std::string sResultFilename = sFilename;
std::string::size_type dot = sResultFilename.rfind('.');
if (dot != std::string::npos)
{
sResultFilename = sResultFilename.substr(0, dot);
}
sResultFilename += "_segmentation.pgm";
if (argc >= 3 && !g_bQATest)
{
sResultFilename = argv[2];
}
// load MRF declaration
int width, height, nLabels;
int *hCue, *vCue, *dataCostArray;
loadMiddleburyMRFData(sFilename, dataCostArray, hCue, vCue, width, height, nLabels);
NPP_ASSERT(nLabels == 2);
std::cout << "Dataset: " << sFilename << std::endl;
std::cout << "Size: " << width << "x" << height << std::endl;
NppiSize size;
size.width = width;
size.height = height;
NppiRect roi;
roi.x=0;
roi.y=0;
roi.width=width;
roi.height=height;
// Setup flow network
int step, transposed_step;
Npp32s *d_source, *d_sink, *d_terminals, *d_left_transposed, *d_right_transposed, *d_top, *d_bottom;
// Setup terminal capacities
d_source = nppiMalloc_32s_C1(width, height, &step);
cudaMemcpy2D(d_source, step, dataCostArray, width * sizeof(int), width*sizeof(int), height, cudaMemcpyHostToDevice);
d_sink = nppiMalloc_32s_C1(width, height, &step);
cudaMemcpy2D(d_sink, step, &dataCostArray[width*height], width * sizeof(int), width*sizeof(int), height, cudaMemcpyHostToDevice);
d_terminals = nppiMalloc_32s_C1(width, height, &step);
nppiSub_32s_C1RSfs(d_sink, step, d_source, step, d_terminals, step, size, 0);
// Setup edge capacities
NppiSize edgeTranposedSize;
edgeTranposedSize.width = height;
edgeTranposedSize.height = width-1;
NppiSize oneRowTranposedSize;
oneRowTranposedSize.width = height;
oneRowTranposedSize.height = 1;
d_right_transposed = nppiMalloc_32s_C1(height, width, &transposed_step);
cudaMemcpy2D(d_right_transposed, transposed_step, hCue, height * sizeof(int), height * sizeof(int), width, cudaMemcpyHostToDevice);
d_left_transposed = nppiMalloc_32s_C1(height, width, &transposed_step);
nppiSet_32s_C1R(0, d_left_transposed, transposed_step, oneRowTranposedSize);
nppiCopy_32s_C1R(d_right_transposed, transposed_step, d_left_transposed + transposed_step/sizeof(int), transposed_step, edgeTranposedSize);
NppiSize edgeSize;
edgeSize.width = width;
edgeSize.height = height-1;
NppiSize oneRowSize;
oneRowSize.width = width;
oneRowSize.height = 1;
d_bottom = nppiMalloc_32s_C1(width, height, &step);
cudaMemcpy2D(d_bottom, step, vCue, width * sizeof(int), width*sizeof(int), height, cudaMemcpyHostToDevice);
d_top = nppiMalloc_32s_C1(width, height, &step);
nppiSet_32s_C1R(0, d_top, step, oneRowSize);
nppiCopy_32s_C1R(d_bottom, step, d_top + step/sizeof(int), step, edgeSize);
// Allocate temp storage for graphcut computation
Npp8u *pBuffer;
int bufferSize;
nppiGraphcutGetSize(size, &bufferSize);
cudaMalloc(&pBuffer, bufferSize);
NppiGraphcutState *pGraphcutState;
nppiGraphcutInitAlloc(size, &pGraphcutState, pBuffer);
// Allocate label storage
npp::ImageNPP_8u_C1 oDeviceDst(width, height);
cudaEvent_t start, stop;
cudaEventCreate(&start);
cudaEventCreate(&stop);
// Compute the graphcut, result is 0 / !=0
cudaEventRecord(start,0);
nppiGraphcut_32s8u(d_terminals, d_left_transposed, d_right_transposed,
d_top, d_bottom, step, transposed_step,
size, oDeviceDst.data(), oDeviceDst.pitch(), pGraphcutState);
cudaEventRecord(stop,0);
cudaEventSynchronize(stop);
float time;
cudaEventElapsedTime(&time, start, stop);
std::cout << "Elapsed Time: " << time << " ms" << std::endl;
// declare a host image object for an 8-bit grayscale image
npp::ImageCPU_8u_C1 oHostAlpha(width, height);
// convert graphcut result to 0/255 alpha image using new nppiCompareC_8u_C1R primitive (CUDA 5.0)
npp::ImageNPP_8u_C1 oDeviceAlpha(width, height);
nppiCompareC_8u_C1R(oDeviceDst.data(), oDeviceDst.pitch(), 0, oDeviceAlpha.data(), oDeviceAlpha.pitch(), size,
NPP_CMP_GREATER);
// and copy the result to host
oDeviceAlpha.copyTo(oHostAlpha.data(), oHostAlpha.pitch());
int E_d, E_s;
std::cout << "Graphcut Cost: " << computeEnergy(E_d, E_s, oHostAlpha.data(), oHostAlpha.pitch(), hCue, vCue, dataCostArray, width, height) << std::endl;
std::cout << "(E_d = " << E_d << ", E_s = " << E_s << ")" << std::endl;
std::cout << "Saving segmentation result as " << sResultFilename << std::endl;
saveImage(sResultFilename, oHostAlpha);
nppiGraphcutFree(pGraphcutState);
cudaFree(pBuffer);
cudaFree(d_top);
cudaFree(d_bottom);
cudaFree(d_left_transposed);
cudaFree(d_right_transposed);
cudaFree(d_source);
cudaFree(d_sink);
cudaFree(d_terminals);
exit(EXIT_SUCCESS);
}
catch (npp::Exception &rException)
{
std::cerr << "Program error! The following exception occurred: \n";
std::cerr << rException << std::endl;
std::cerr << "Aborting." << std::endl;
exit(EXIT_FAILURE);
}
catch (...)
{
std::cerr << "Program error! An unknow type of exception occurred. \n";
std::cerr << "Aborting." << std::endl;
exit(EXIT_FAILURE);
}
return 0;
}
int main()
{
int timer = 0,step,solution=0, useNew, accepted;
float temp = INIT_TEMP;
memberType current,working,best;
FILE *fp;
fp=fopen("stats.txt", "w");
srand(time(NULL));
initSolution(¤t);
computeEnergy(¤t);
best.energy=100.0;
copySolution(&working, ¤t);
while( temp > FIN_TEMP )
{
printf("\n Temperature = %f", temp);
accepted = 0;
/* Monte Carlo step*/
for( step = 0; step < STEPS; step++);
{
useNew=0;
tweakSolution(&working);
computeEnergy(&working);
if(working.energy <= current.energy)
{
useNew = 1;
}
else
{
float test = rand() % 1;
float delta = working.energy - current.energy;
float calc = exp(-delta/temp);
if(calc > test)
{
accepted++;
useNew = 1;
}
}
}
if(useNew)
{
useNew = 0;
copySolution(¤t, &working);
if(current.energy < best.energy)
{
copySolution(&best, ¤t);
solution = 1;
}
else
{
copySolution(&working, ¤t);
}
}
fprintf(fp, "%d %f %f %d \n", timer++, temp, best.energy, accepted);
printf("Best Energy = %f\n", best.energy);
temp *= ALPHA;
}
fclose(fp);
if(solution)
{
emitSolution(&best);
}
return 0;
}
void Evnav::route(EvnavRequest &req, QJsonObject &json)
{
Trip trip;
Graph g;
VertexId srcId = -1;
VertexId dstId = -2;
double SOC_dyn = req.m_SOC_max - req.m_SOC_min;
double batt_act = req.m_battery * req.m_SOC_act; // kWh
double batt_dyn = req.m_battery * SOC_dyn;
if (computeTrip(req.m_src, req.m_dst, trip) == Status::Ok) {
double e = computeEnergy(trip, req.m_efficiency);
double e_otw = 0; // kWh
if (e > batt_act) {
e_otw = e - batt_act;
}
qDebug() << "energy required : " << e << "kWh";
qDebug() << "energy start : " << batt_act << "kWh";
qDebug() << "energy on the way: " << e_otw << "kWh";
if (e < batt_act) {
qDebug() << "reaching destination without charging";
// FIXME: collect result processing
json["code"] = "Ok";
json["message"] = "reachable";
QJsonObject summary;
summary["distance"] = trip.dist_m;
summary["duration"] = trip.time_s;
summary["energy"] = e;
summary["driving_duration"] = trip.time_s;
summary["charging_duration"] = 0;
summary["charging_cost"] = 0;
json["route_summary"] = summary;
json["charging_steps"] = QJsonArray{};
return;
} else {
int min_stops = std::ceil(e_otw / batt_dyn);
double charging_time = computeChargingTime(e_otw, req.m_power_avg);
qDebug() << "charging min_stops:" << min_stops;
qDebug() << "charging min_time :" << formatTime(charging_time);
}
}
makeGraph(g, req, srcId, dstId);
qDebug() << "graph size:" << g.E();
// TODO: write the graph as Json
ShortestPath sp(g, srcId);
if (sp.hasPathTo(dstId)) {
QVector<Edge> path = sp.pathTo(dstId).toVector();
write(path, json);
} else {
json["code"] = "Ok";
json["message"] = "unreachable";
qDebug() << "cannot reach destination with this electric car";
}
}
