SimonBlanke /
Gradient-Free-Optimizers
| Total Complexity | 3 |
| Total Lines | 33 |
| Duplicated Lines | 0 % |
| Changes | 0 | ||
| 1 | # Author: Simon Blanke |
||
| 2 | # Email: [email protected] |
||
| 3 | # License: MIT License |
||
| 4 | |||
| 5 | |||
| 6 | import numpy as np |
||
| 7 | |||
| 8 | from ..local_opt import StochasticHillClimbingOptimizer |
||
| 9 | |||
| 10 | |||
| 11 | class SimulatedAnnealingOptimizer(StochasticHillClimbingOptimizer): |
||
| 12 | name = "Simulated Annealing" |
||
| 13 | _name_ = "simulated_annealing" |
||
| 14 | __name__ = "SimulatedAnnealingOptimizer" |
||
| 15 | |||
| 16 | optimizer_type = "local" |
||
| 17 | computationally_expensive = False |
||
| 18 | |||
| 19 | def __init__(self, *args, annealing_rate=0.97, start_temp=1, **kwargs): |
||
| 20 | super().__init__(*args, **kwargs) |
||
| 21 | |||
| 22 | self.annealing_rate = annealing_rate |
||
| 23 | self.start_temp = start_temp |
||
| 24 | self.temp = start_temp |
||
| 25 | |||
| 26 | def _p_accept_default(self): |
||
| 27 | # the 'minus' is omitted, because we maximize a score |
||
| 28 | return np.exp(self._exponent) |
||
| 29 | |||
| 30 | def evaluate(self, score_new): |
||
| 31 | StochasticHillClimbingOptimizer.evaluate(self, score_new) |
||
| 32 | self.temp *= self.annealing_rate |
||
| 33 |
