void GeneticPopulation::crossover(
const Weights& parent1,
const Weights& parent2,
Weights& child1,
Weights& child2) const {
assert(parent1.size() == parent2.size());
if (randomReal(0, 1) > crossoverRate || parent1 == parent2) {
child1 = parent1;
child2 = parent2;
return;
}
unsigned crossoverPoint = static_cast<unsigned>(randomInt(0, parent1.size()));
child1.clear();
child2.clear();
//create the offspring
for (unsigned i = 0; i < crossoverPoint; ++i) {
child1.push_back(parent1[i]);
child2.push_back(parent2[i]);
}
for (unsigned i = crossoverPoint; i < parent1.size(); ++i) {
child1.push_back(parent2[i]);
child2.push_back(parent1[i]);
}
}
void printHeader(const Alphabet& alphabet, const Weights& weights, Options::Mode mode) {
cout <<
"# " PROGRAM_NAME " " PROGRAM_VERSION "\n"
"# Copyright 2006 Informatics for Mass Spectrometry group\n"
"# at Bielefeld University\n"
"#\n"
"# http://BiBiServ.TechFak.Uni-Bielefeld.DE/decomp/\n"
"#\n"
"# alphabet (character, weight):\n";
for (Weights::size_type i = 0; i < weights.size(); ++i) {
cout
<< "#\t" << alphabet.getName(i) << "\t"
<< weights[i] << '\n';
}
cout << "#\n# Problem to solve: ";
switch (mode) {
case Options::GETNUMBER:
cout << "Get number of decompositions";
break;
case Options::FINDALL:
cout << "Find all decompositions";
break;
case Options::FINDONE:
cout << "Find one decomposition";
break;
case Options::ISDECOMPOSABLE:
cout << "Is decomposable";
break;
}
cout << "\n#\n";
}
ReturnCode computeMeanSigma(Result &mean, Result &sigma, const Values &values, const Weights &weights)
{
MSS_BEGIN(ReturnCode);
MSS(!values.empty());
MSS(values.size() == weights.size());
//Compute mean and the sum of the weights
mean = 0.0;
Result sumW = 0.0;
typename Values::const_iterator value = values.begin();
for (auto weight: weights)
{
mean += *(value++)*weight;
sumW += weight;
}
mean /= sumW;
//Compute sigma
sigma = 0.0;
typename Weights::const_iterator weight = weights.begin();
for (auto value: values)
sigma += *(weight++)*(value-mean)*(value-mean);
sigma = ::sqrt(sigma/sumW);
MSS_END();
}
void GameManager::controlCar() {
Weights outputs = callNeuralNetwork();
assert(outputs.size() == 3);
Car& car = model.getCar();
float throttleOutput = clamp((2.f/3.f)*outputs[0] + (2.f/3.f), 0.f, 1.f);
float brakeOutput = clamp((2.f/3.f)*outputs[1] + (1.f/3.f), 0.f, 1.f);
float turnLevelOutput = outputs[2];
car.setThrottle(throttleOutput);
car.setBrake(brakeOutput);
car.setTurnLevel(turnLevelOutput);
}
ReturnCode computeMean(Mean &mean, const Values &values, const Weights &weights)
{
MSS_BEGIN(ReturnCode);
MSS(!values.empty());
MSS(values.size() == weights.size());
mean = 0.0;
Mean sumW = 0.0;
typename Values::const_iterator value = values.begin();
for (auto weight: weights)
{
mean += *(value++)*weight;
sumW += weight;
}
mean /= sumW;
MSS_END();
}
int random_weighted_choice(const Weights& w)
{
using T = typename Weights::value_type;
auto n = w.size();
TBLIS_ASSERT(n > 0);
T s = 0;
for (unsigned i = 0;i < n;i++)
{
TBLIS_ASSERT(w[i] >= 0);
s += w[i];
}
T c = random_number(s);
for (unsigned i = 0;i < n;i++)
{
if (c < w[i]) return i;
c -= w[i];
}
return n-1;
}
{
TEST_EQUAL(ptr->getAlphabetMass(i), copy.getAlphabetMass(i))
TEST_EQUAL(ptr->getWeight(i), copy.getWeight(i))
TEST_EQUAL((*ptr)[i], copy[i])
}
}
END_SECTION
START_SECTION((Weights& operator=(const Weights &weights_)))
{
Weights copy;
copy = *ptr;
TEST_NOT_EQUAL(©, null_ptr)
// test equality of copy and ptr
ABORT_IF(ptr->size()!=copy.size())
for(Size i = 0 ; i < ptr->size() ; ++i)
{
TEST_EQUAL(ptr->getAlphabetMass(i), copy.getAlphabetMass(i))
TEST_EQUAL(ptr->getWeight(i), copy.getWeight(i))
TEST_EQUAL((*ptr)[i], copy[i])
}
}
END_SECTION
START_SECTION((size_type size() const ))
{
TEST_EQUAL(ptr->size(), 5)
Weights w;
TEST_EQUAL(w.size(),0)
