evomo.problems.constrained.sdc

Classes

SDC	Scalable high-dimensional decicsion constraint benchamrk
SDC1	Scalable high-dimensional decicsion constraint benchamrk
SDC2	Scalable high-dimensional decicsion constraint benchamrk
SDC3	Scalable high-dimensional decicsion constraint benchamrk
SDC4	Scalable high-dimensional decicsion constraint benchamrk
SDC5	Scalable high-dimensional decicsion constraint benchamrk
SDC6	Scalable high-dimensional decicsion constraint benchamrk
SDC7	Scalable high-dimensional decicsion constraint benchamrk
SDC8	Scalable high-dimensional decicsion constraint benchamrk
SDC9	Scalable high-dimensional decicsion constraint benchamrk
SDC10	Scalable high-dimensional decicsion constraint benchamrk
SDC11	Scalable high-dimensional decicsion constraint benchamrk
SDC12	Scalable high-dimensional decicsion constraint benchamrk
SDC13	Scalable high-dimensional decicsion constraint benchamrk
SDC14	Scalable high-dimensional decicsion constraint benchamrk
SDC15	Scalable high-dimensional decicsion constraint benchamrk

Functions

information(→ list)
CEC_2006_information(problem, *[, device, dtype])
CEC_2006_fitness(→ Tuple[torch.Tensor, torch.Tensor])
Distance_function(→ torch.Tensor)

Module Contents

evomo.problems.constrained.sdc.information(index: int) → list

evomo.problems.constrained.sdc.CEC_2006_information(problem: int, *, device: torch.device | None = None, dtype: torch.dtype | None = None)

evomo.problems.constrained.sdc.CEC_2006_fitness(P: torch.Tensor, problem: int, aaa: torch.Tensor, optimal: torch.Tensor) → Tuple[torch.Tensor, torch.Tensor]

evomo.problems.constrained.sdc.Distance_function(x: torch.Tensor, problem: int) → torch.Tensor

class evomo.problems.constrained.sdc.SDC(info_idx: int, d: int = 30, m: int = 2, ref_num: int = 1000, device: torch.device | None = None, dtype: torch.dtype | None = None, **kwargs)

Bases: evomo.problems.constrained.base.CMOP

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

m = 2

ref_num = 1000

info

lu = None

high_D_C = None

aaa = None

optial_f

max_D_con

fn(X: torch.Tensor) → torch.Tensor

pf()

class evomo.problems.constrained.sdc.SDC1(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

class evomo.problems.constrained.sdc.SDC2(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

class evomo.problems.constrained.sdc.SDC3(info_idx=3, d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

pf()

class evomo.problems.constrained.sdc.SDC4(info_idx=4, d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

pf()

class evomo.problems.constrained.sdc.SDC5(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC3

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC6(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC4

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC7(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC8(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC9(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC3

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC10(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC4

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC11(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC3

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC12(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC4

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC13(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC4

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC14(d: int = 30, m: int = 2, ref_num: int = 1000, **kwargs)

Bases: SDC3

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

class evomo.problems.constrained.sdc.SDC15(d: int = 30, m: int = 3, ref_num: int = 1000, **kwargs)

Bases: SDC

Scalable high-dimensional decicsion constraint benchamrk

K. Qiao, J. Liang, K. Yu, C. Yue, H. Lin, D. Zhang, and B. Qu. Evolutionary constrained multiobjective optimization: scalable high-dimensional constraint benchmarks and algorithm. IEEE Transactions on Evolutionary Computation, 2024, 28(4): 965-979.

param: THETA_ a CEC_Problem: index of high-dimension constraint function Distance_problem: index of distance function HCT: rate of unconstrained distance function variables that are used for high-dimensional constraint function HCT_type: type of transformation function DCT: rate of unconstrained distance function variables that are used for variable linkage DCT_type: type of variable linkage function Shape_problem: tyep of shape function b: a parameter of shape function that is used to control the overlap degree between CPF and UPF lu: upper and lower bounds of high-dimensional constraint function high_D_C: dimensions of high-dimensional constraint function aaa: used for high-dimensional constraint function optial_f: optimal objective value of high-dimensional constraints q1_upper: upper bound of shape function q1_lower:lower bound of shape function h: objective function value of high-dimensional constraint function HC: constraint function value of high-dimensional constraint function max_D_con: maximal number of variables that are used for transformation operator

ref_num = 1000

fn(X: torch.Tensor) → torch.Tensor

Return: [F1, F2, F3, c1, c2, HC] with shape (N, 6) - F*: 3 objectives (M=3) - c1,c2: shape constraints from SDC15 - HC: aggregated high-dimensional constraint violation from CEC_2006