Aggregation Functions #

File position: /EMOC/src/core/utility.h and /EMOC/src/core/utility.cpp

There are several aggregation functions for decomposition:

double CalWeightedSum(Individual *ind, double *weight_vector, double *ideal_point, int obj_num)

Use weighted sum method to decompose the individual ind based on weight vector weight_vector.

Parameter:
	ind: Individual, default=None* The pointer to the individual which will be decomposed. weight_vector: double, default=None* The weight vector for decomposition. It's a 1D array of double. obj_num: int, default=None The number of objectives.
Returns:
	weighted_sum: double The decomposition value with the weighted sum method.

double CalInverseChebycheff(Individual *ind, double *weight_vector, double *ideal_point, int obj_num)

Use inverse chebycheff method to decompose the individual ind based on weight vector weight_vector.

Parameter:
	ind: Individual, default=None* The pointer to the individual which will be decomposed. weight_vector: double, default=None* The weight vector for decomposition. It's a 1D array of double. ideal_point: double, default=None* The ideal point of current population. It's a 1D array of double. obj_num: int, default=None The number of objectives.
Returns:
	inverse_chebycheff: double The decomposition value with the inverse chebycheff method.

double CalPBI(Individual *ind, double *weight_vector, double *ideal_point, int obj_num, double theta)

Use penalty-based boundary intersection method to decompose the individual ind based on weight vector weight_vector.

Parameter:
	ind: Individual, default=None* The pointer to the individual which will be decomposed. weight_vector: double, default=None* The weight vector for decomposition. It's a 1D array of double. ideal_point: double, default=None* The ideal point of current population. It's a 1D array of double. obj_num: int, default=None The number of objectives. theta: double, default=None The parameter of penalty-based boundary intersection method.
Returns:
	pbi: double The decomposition value with the penalty-based boundary intersection method.