summaryrefslogtreecommitdiff
path: root/src/sample_selection.cc
diff options
context:
space:
mode:
authorAndré Nusser <andre.nusser@googlemail.com>2019-03-08 01:21:24 +0100
committerAndré Nusser <andre.nusser@googlemail.com>2019-05-11 14:54:51 +0200
commitee41b02f78a81cb96365c56f7a8dd69d7f6e7d72 (patch)
tree917f888174cf129c21ce945b34f431c0cb5d5eb6 /src/sample_selection.cc
parentdcff4b6914d07c6dcff9cbbdc93dc0f4dc146d44 (diff)
Optimized version of new sample selection.
Diffstat (limited to 'src/sample_selection.cc')
-rw-r--r--src/sample_selection.cc51
1 files changed, 44 insertions, 7 deletions
diff --git a/src/sample_selection.cc b/src/sample_selection.cc
index 8616e09..8319017 100644
--- a/src/sample_selection.cc
+++ b/src/sample_selection.cc
@@ -32,6 +32,8 @@
#include "random.h"
#include "settings.h"
+#include <algorithm>
+
namespace
{
@@ -181,6 +183,7 @@ const Sample* SampleSelection::getObjective(level_t level, std::size_t pos)
float distance_opt = 0.;
float recent_opt = 0.;
+ // TODO: check how much sense all of this actually makes
// Select normal distributed value between
// (stddev/2) and (power_span-stddev/2)
float lvl = rand.normalDistribution(mean, stddev);
@@ -200,26 +203,60 @@ const Sample* SampleSelection::getObjective(level_t level, std::size_t pos)
// TODO: start with most promising power value and then stop when reaching far values
// which cannot become opt anymore
- for (std::size_t i = 0; i < samples.size(); ++i)
+ // TODO: can we expect the powerlist to be sorted? If not, add to finalise of powerlist.
+ // TODO: clean up this mess
+ auto closest_it = std::lower_bound(samples.begin(), samples.end(), level);
+ std::size_t up_index = std::distance(samples.begin(), closest_it);
+ std::size_t down_index = (up_index == 0 ? 0 : up_index - 1);
+ float up_value_lb = (up_index < samples.size() ? alpha*pow2(samples[up_index].power-lvl) : std::numeric_limits<float>::max());
+ float down_value_lb = (up_index != 0 ? alpha*pow2(samples[down_index].power-lvl) : std::numeric_limits<float>::max());
+ do
{
- auto const& item = samples[i];
+ std::size_t current_index;
+ if (up_value_lb < down_value_lb)
+ {
+ current_index = up_index;
+ if (up_index != samples.size()-1)
+ {
+ ++up_index;
+ up_value_lb = alpha*pow2(samples[up_index].power-lvl);
+ }
+ else
+ {
+ up_value_lb = std::numeric_limits<float>::max();
+ }
+ }
+ else
+ {
+ current_index = down_index;
+ if (down_index != 0)
+ {
+ --down_index;
+ down_value_lb = alpha*pow2(samples[down_index].power-lvl);
+ }
+ else
+ {
+ down_value_lb = std::numeric_limits<float>::max();
+ }
+ }
- // compute objective function value
auto random = rand.floatInRange(0.,1.);
- auto distance = item.power - lvl;
- auto recent = (float)settings.samplerate/std::max<std::size_t>(pos - last[i], 1);
+ auto distance = samples[current_index].power - lvl;
+ auto recent = (float)settings.samplerate/std::max<std::size_t>(pos - last[current_index], 1);
auto value = alpha*pow2(distance) + beta*pow2(recent) + gamma*random;
if (value < value_opt)
{
- index_opt = i;
- power_opt = item.power;
+ index_opt = current_index;
+ power_opt = samples[current_index].power;
value_opt = value;
random_opt = random;
distance_opt = distance;
recent_opt = recent;
}
+ --current_index;
}
+ while (up_value_lb <= value_opt || down_value_lb <= value_opt);
DEBUG(rand, "Chose sample with index: %d, value: %f, power %f, random: %f, distance: %f, recent: %f", (int)index_opt, value_opt, power_opt, random_opt, distance_opt, recent_opt);