Similarity-based Grouping Genetic Algorithm Parameters. More...
#include <sgga.h>
Public Member Functions | |
RParamsSG (void) | |
void | InitObjs (size_t highid, size_t nb) |
bool | IsSocial (size_t id) |
void | SetSocial (size_t id, bool social) |
void | AddCloseObject (size_t id, size_t closeid, double ratio) |
template<class cGroup , class cGroups > | |
void | FindBestGroup (size_t id, cGroups *groups, cGroup *&grp) |
Public Attributes | |
RContainer< RObjRefSG, true, false > | Objs |
size_t | PopSize |
size_t | MaxGen |
bool | Step |
size_t | StepGen |
double | MinSimLevel |
double | MinAgreement |
double | MinDisagreement |
size_t | MaxKMeans |
double | Convergence |
size_t | NbDivChromo |
RParam * | ParamsSim |
RParam * | ParamsAgreement |
RParam * | ParamsDisagreement |
bool | LocalOptimisation |
bool | Optimisation |
bool | Incremental |
size_t | NbMinObjs |
size_t | NbMaxObjs |
bool | AllMinSim |
Detailed Description
Similarity-based Grouping Genetic Algorithm Parameters.
The RParamsSG represents all the parameter used in the similarity-based grouping genetic algorithm.
Constructor & Destructor Documentation
RParamsSG | ( | void | ) |
Constructor.
Member Function Documentation
void InitObjs | ( | size_t | highid, |
size_t | nb | ||
) |
Allocate the object references.
- Parameters
-
highid Highest identifier. nb Supposed number of closest objects.
bool IsSocial | ( | size_t | id | ) |
Look if a given object is social.
- Parameters
-
id Identifier of the object.
- Returns
- a boolean
void SetSocial | ( | size_t | id, |
bool | social | ||
) |
Set if a given object is social or not.
- Parameters
-
id Identifier of the object. social Social object ?
void AddCloseObject | ( | size_t | id, |
size_t | closeid, | ||
double | ratio | ||
) |
Add a close object to an object.
- Parameters
-
id Identifier of the object. closeid Identifier of a close object. ratio Ratio representing the closeness.
void FindBestGroup | ( | size_t | id, |
cGroups * | groups, | ||
cGroup *& | grp | ||
) |
Find the best group for the object based on the current clustering and the ratio.
- Parameters
-
id Identifier of the object. groups Clustering. grp Variable that will hold the pointer to the group (or 0 if no group was found).
Member Data Documentation
RContainer<RObjRefSG,true,false> Objs |
All closed objects.
size_t PopSize |
Size of the population of the GA.
size_t MaxGen |
Maximal number of generation to run.
bool Step |
Is the GA in step mode?
size_t StepGen |
Number of generation for each step.
double MinSimLevel |
Minimum threshold for similarity between the objects.
double MinAgreement |
Minimum threshold for the agreement ratio.
double MinDisagreement |
Minimum threshold for the disagreement ratio.
size_t MaxKMeans |
Maximal Number of k-Means.
double Convergence |
Convergence ratio between two k-Means iterations.
size_t NbDivChromo |
Number of derived chromosomes must be constructed for the optimization.
RParam* ParamsSim |
Parameter for the criterion "Similarity".
RParam* ParamsAgreement |
Parameter for the criterion "Agreement".
RParam* ParamsDisagreement |
Parameter for the criterion "Disagreement".
bool LocalOptimisation |
Perform a local optimization.
bool Optimisation |
Perform an optimization.
bool Incremental |
Incremental mode. The GCA constructs the chromosomes from existing solutions.
size_t NbMinObjs |
Minimum number of objects per groups.
size_t NbMaxObjs |
Maximum number of objects per group.
bool AllMinSim |
All the pairs of objects of a group must respect the minimum of similarity constraint.