SimonBlanke /
Gradient-Free-Optimizers
| Total Complexity | 1 |
| Total Lines | 40 |
| Duplicated Lines | 0 % |
| Changes | 0 | ||
| 1 | # Author: Simon Blanke |
||
| 2 | # Email: [email protected] |
||
| 3 | # License: MIT License |
||
| 4 | |||
| 5 | |||
| 6 | from .exp_imp_based_opt import ExpectedImprovementBasedOptimization |
||
| 7 | |||
| 8 | from .surrogate_models import ( |
||
| 9 | GPR_linear, |
||
| 10 | GPR, |
||
| 11 | ) |
||
| 12 | |||
| 13 | gaussian_process = {"gp_nonlinear": GPR(), "gp_linear": GPR_linear()} |
||
| 14 | |||
| 15 | |||
| 16 | class BayesianOptimizer(ExpectedImprovementBasedOptimization): |
||
| 17 | name = "Bayesian Optimization" |
||
| 18 | |||
| 19 | def __init__( |
||
| 20 | self, |
||
| 21 | *args, |
||
| 22 | gpr=gaussian_process["gp_nonlinear"], |
||
| 23 | xi=0.03, |
||
| 24 | warm_start_smbo=None, |
||
| 25 | max_sample_size=10000000, |
||
| 26 | sampling={"random": 1000000}, |
||
| 27 | warnings=100000000, |
||
| 28 | **kwargs, |
||
| 29 | ): |
||
| 30 | super().__init__(*args, **kwargs) |
||
| 31 | self.gpr = gpr |
||
| 32 | self.regr = gpr |
||
| 33 | self.xi = xi |
||
| 34 | self.warm_start_smbo = warm_start_smbo |
||
| 35 | self.max_sample_size = max_sample_size |
||
| 36 | self.sampling = sampling |
||
| 37 | self.warnings = warnings |
||
| 38 | |||
| 39 | self.init_warm_start_smbo() |
||
| 40 |
