|
GEGELATI
|
#include <mapElitesDefaultDescriptors.h>
Public Member Functions | |
| virtual void | initDescriptor (const TPG::TPGGraph &graph, const Learn::LearningEnvironment &learningEnvironment) override |
| Specialisation of initDescriptor. | |
| virtual std::string | getName () const override |
| Specialisation of getName. | |
| virtual void | extractMetricsStep (std::vector< double > &metrics, const TPG::TPGVertex *agent, std::vector< double > actionValues, const Learn::LearningEnvironment &learningEnvironment) const override |
| Specialisation of extractMetricsStep. | |
| virtual void | extractMetricsEpisode (std::vector< double > &metrics, const TPG::TPGVertex *agent, size_t nbStepsExecuted, const Learn::LearningEnvironment &learningEnvironment) const override |
| Specialisation of extractMetricsEpisode. | |
Public Member Functions inherited from Selector::MapElites::MapElitesDescriptor | |
| virtual bool | isInit () const |
| method that return if the initDescriptor method has been executed | |
| virtual std::pair< double, double > | getMinAndMaxRange () const |
| Get the min and max range of the descriptor. | |
| virtual size_t | getNbDescriptors () const |
| Get the number of the descriptor. | |
Additional Inherited Members | |
Protected Attributes inherited from Selector::MapElites::MapElitesDescriptor | |
| size_t | nbDescriptors |
| Number of descriptor. | |
| double | minRange |
| Minimum value in the archive. | |
| double | maxRange |
| Maximum value in the archive. | |
| bool | init = false |
Class for ActionValues descriptor.
This descriptor is designed for continuous environments (an update could be made to be used on discrete environments too but it might not be useful).
It computes the average of the absolute of the action values over an episode.
|
overridevirtual |
Specialisation of extractMetricsEpisode.
Reimplemented from Selector::MapElites::MapElitesDescriptor.
|
overridevirtual |
Specialisation of extractMetricsStep.
Reimplemented from Selector::MapElites::MapElitesDescriptor.
|
overridevirtual |
Specialisation of getName.
Implements Selector::MapElites::MapElitesDescriptor.
|
overridevirtual |
Specialisation of initDescriptor.
Copyright or © or Copr. IETR/INSA - Rennes (2025) :
Quentin Vacher qvach.nosp@m.er@i.nosp@m.nsa-r.nosp@m.enne.nosp@m.s.fr (2025)
GEGELATI is an open-source reinforcement learning framework for training artificial intelligence based on Tangled Program Graphs (TPGs).
This software is governed by the CeCILL-C license under French law and abiding by the rules of distribution of free software. You can use, modify and/ or redistribute the software under the terms of the CeCILL-C license as circulated by CEA, CNRS and INRIA at the following URL "http://www.cecill.info".
As a counterpart to the access to the source code and rights to copy, modify and redistribute granted by the license, users are provided only with a limited warranty and the software's author, the holder of the economic rights, and the successive licensors have only limited liability.
In this respect, the user's attention is drawn to the risks associated with loading, using, modifying and/or developing or reproducing the software by the user in light of its specific status of free software, that may mean that it is complicated to manipulate, and that also therefore means that it is reserved for developers and experienced professionals having in-depth computer knowledge. Users are therefore encouraged to load and test the software's suitability as regards their requirements in conditions enabling the security of their systems and/or data to be ensured and, more generally, to use and operate it in the same conditions as regards security.
The fact that you are presently reading this means that you have had knowledge of the CeCILL-C license and that you accept its terms.
Implements Selector::MapElites::MapElitesDescriptor.