This package offers some tools on randomized optimization, namely preset problems, utilities to create new problems and predefined optimization algorithms providing limited customization in terms of parameters.

• HillClimber
• RandomizedRestarts (can be used in combination with any optimizer)
• Simulated Annealing
• Genetic Algorithm
• MIMIC

• Alternating bits (1010101)
• Bulls & cows
• Knapsack
• Guess my bits
• Neural Network weight optimization

Below is a generic usage for problems and solvers which is applicable to all solver and problem classes offered under this package:

import randopts as rd

>>> problem = rd.problem.AlternateProblem(nbits=7)
>>> solver = rd.solver.HillClimber()
>>> solver.fit(problem)  # parses parameters such as required bits
>>> solver.ara(problem)  # attempts to solve the problem
'1010101'

>>> problem.query_cost_td
45.0

>>> solver.maks
6

>>> solver.solution_arr
[3,
 3,
...
 5,
 6]

>>> solver.solution_time
0.001