Simulated Anealing pytorch

This is an pytorch Optimizer() using Simulating Annealing Algorithm to find the target solution.

# Code Structure
.
├── LICENSE
├── Readme.md
├── Simulated_Annealing_Optimizer.py  # SimulatedAnealling(optim.Optimizer)
├── demo.py                           # Demo using Simulated Annealing to solve a question
└── fig 
    └── test_animation.gif

We can use:

from Simulated_Annealing_Optimizer import SimulatedAnealling

to import the optimizer. And we can initialize the optimizer by

self.optimizer=SimulatedAnealling(params=model.parameters(),init_temp=args.init_temp,cooling=args.cooling
                                          ,lr=args.lr,model=self.model,loss_fn=self.loss_fn,y=self.y)

Algorithm

The procedure of the algorithm can be described as follow：

Demo

demo.py provides a little demo describes how optimizer could be used. We want to find the minimum of a polynomial function Followed is a .gif shows how minimum is found by Simulated Annealing Algorithm.