SedSat3 1.1.6
Sediment Source Apportionment Tool - Advanced statistical methods for environmental pollution research
Loading...
Searching...
No Matches
Public Attributes | List of all members
GA_Tweaking_parameters Struct Reference

Configuration parameters for Genetic Algorithm optimization. More...

#include <GA/GA.h>

Collaboration diagram for GA_Tweaking_parameters:
Collaboration graph

Public Attributes

int totnumparams
 Total number of parameters in the optimization problem.
 
int maxpop = 2
 Maximum population size (number of individuals)
 
int nParam
 Number of parameters (may be subset of totnumparams)
 
int numenhancements
 Number of enhancement iterations.
 
int num_enh
 Current enhancement iteration counter.
 
int nGen
 Number of generations to evolve.
 
int cross_over_type
 Crossover operator type.
 
int no_bins
 Number of bins for fitness distribution.
 
bool sens_out = false
 Output sensitivity analysis results.
 
bool RCGA = false
 Use Real-Coded Genetic Algorithm.
 
bool readfromgafile
 Read initial population from file.
 
bool Steepest_Descent = false
 Apply steepest descent local search.
 
double N
 Selection pressure parameter for rank-based selection.
 
double pcross = 1
 Crossover probability.
 
double pmute = 0.02
 Mutation probability per gene.
 
double exponentcoeff = 1
 Exponent coefficient for fitness scaling.
 
double shakescale = 0.05
 Scale factor for shake/perturbation operation.
 
double shakescalered = 0.75
 Reduction factor for shake scale over generations.
 
char fitnesstype
 Fitness function type identifier.
 

Detailed Description

Configuration parameters for Genetic Algorithm optimization.

Contains all tuning parameters and settings that control the behavior of the genetic algorithm, including population size, genetic operators, and convergence criteria.

Definition at line 26 of file GA.h.

Member Data Documentation

◆ cross_over_type

int GA_Tweaking_parameters::cross_over_type

Crossover operator type.

Selects which crossover strategy to use for combining parent solutions:

  • Type 0: Single-point crossover
  • Type 1: Two-point crossover
  • Type 2: Uniform crossover
  • Other: Problem-specific operators

Definition at line 85 of file GA.h.

◆ exponentcoeff

double GA_Tweaking_parameters::exponentcoeff = 1

Exponent coefficient for fitness scaling.

Used in exponential fitness transformation to control selection pressure: scaled_fitness = exp(exponentcoeff × raw_fitness)

Default: 1.0

Definition at line 177 of file GA.h.

◆ fitnesstype

char GA_Tweaking_parameters::fitnesstype

Fitness function type identifier.

Character code specifying which objective function to use:

  • 'L': Least squares error
  • 'M': Maximum likelihood
  • 'C': Custom fitness function
  • Other: Problem-specific functions

Definition at line 208 of file GA.h.

◆ maxpop

int GA_Tweaking_parameters::maxpop = 2

Maximum population size (number of individuals)

Number of candidate solutions maintained in each generation. Larger populations explore more broadly but require more evaluations.

Typical: 50-200 individuals Default: 2 (likely intended as multiplier, not actual size)

Definition at line 45 of file GA.h.

◆ N

double GA_Tweaking_parameters::N

Selection pressure parameter for rank-based selection.

Controls intensity of selection pressure in rank-based fitness assignment. Higher N gives stronger preference to better individuals.

Typical: 1.5-2.0

Definition at line 145 of file GA.h.

◆ nGen

int GA_Tweaking_parameters::nGen

Number of generations to evolve.

Total iterations of selection, crossover, and mutation. More generations allow better convergence but take longer.

Typical: 100-1000 generations

Definition at line 74 of file GA.h.

◆ no_bins

int GA_Tweaking_parameters::no_bins

Number of bins for fitness distribution.

Used in fitness proportionate selection to discretize fitness values for roulette wheel selection.

Definition at line 93 of file GA.h.

◆ nParam

int GA_Tweaking_parameters::nParam

Number of parameters (may be subset of totnumparams)

Number of active parameters being optimized (non-fixed parameters).

Definition at line 52 of file GA.h.

◆ num_enh

int GA_Tweaking_parameters::num_enh

Current enhancement iteration counter.

Definition at line 64 of file GA.h.

◆ numenhancements

int GA_Tweaking_parameters::numenhancements

Number of enhancement iterations.

Additional optimization refinements applied to best solutions.

Definition at line 59 of file GA.h.

◆ pcross

double GA_Tweaking_parameters::pcross = 1

Crossover probability.

Probability that two selected parents undergo crossover. If crossover doesn't occur, parents are copied to offspring unchanged.

Typical: 0.6-0.9 Default: 1.0 (always crossover)

Definition at line 156 of file GA.h.

◆ pmute

double GA_Tweaking_parameters::pmute = 0.02

Mutation probability per gene.

Probability that each parameter value is randomly perturbed. Maintains diversity and enables exploration of new regions.

Typical: 0.001-0.01 for binary, 0.01-0.1 for real-coded Default: 0.02 (2%)

Definition at line 167 of file GA.h.

◆ RCGA

bool GA_Tweaking_parameters::RCGA = false

Use Real-Coded Genetic Algorithm.

If true, uses continuous (real-valued) representation and operators. If false, uses binary encoding and operators.

RCGA is generally superior for continuous optimization problems.

Default: false

Definition at line 115 of file GA.h.

◆ readfromgafile

bool GA_Tweaking_parameters::readfromgafile

Read initial population from file.

If true, loads starting population from file instead of random initialization. Useful for warm-starting optimization or continuing previous runs.

Default: false

Definition at line 125 of file GA.h.

◆ sens_out

bool GA_Tweaking_parameters::sens_out = false

Output sensitivity analysis results.

If true, calculates and saves parameter sensitivity information during optimization.

Default: false

Definition at line 103 of file GA.h.

◆ shakescale

double GA_Tweaking_parameters::shakescale = 0.05

Scale factor for shake/perturbation operation.

Controls magnitude of random perturbations in shake operation, specified as fraction of parameter range.

Default: 0.05 (5% of range)

Definition at line 187 of file GA.h.

◆ shakescalered

double GA_Tweaking_parameters::shakescalered = 0.75

Reduction factor for shake scale over generations.

Shake scale is multiplied by this factor each generation to gradually reduce perturbation magnitude (simulated annealing effect).

Default: 0.75 (25% reduction per generation)

Definition at line 197 of file GA.h.

◆ Steepest_Descent

bool GA_Tweaking_parameters::Steepest_Descent = false

Apply steepest descent local search.

If true, applies gradient-based refinement to best individuals to accelerate convergence near optima. Creates hybrid GA/local-search algorithm.

Default: false

Definition at line 135 of file GA.h.

◆ totnumparams

int GA_Tweaking_parameters::totnumparams

Total number of parameters in the optimization problem.

Dimensionality of the parameter space. For sediment fingerprinting, typically equals the number of sources to be apportioned.

Definition at line 34 of file GA.h.

The documentation for this struct was generated from the following file: