Coverage for moptipy / api / mo_algorithm.py: 91%

1""" 

2The base classes for multi-objective optimization algorithms. 

3 

4A multi-objective optimization algorithm is an optimization algorithm that can 

5optimize multiple, possibly conflicting, objective functions at once. All such 

6algorithms inherit from :class:`~moptipy.api.mo_algorithm.MOAlgorithm`. When 

7developing or implementing new algorithms, you would still follow the concepts 

8discussed in module :mod:`~moptipy.api.algorithm`. 

9 

10If you implement a new multi-objective algorithm, you can test it via the 

11pre-defined unit test routine 

12:func:`~moptipy.tests.mo_algorithm.validate_mo_algorithm`. 

13""" 

14from typing import cast 

15 

16from pycommons.types import type_error 

17 

18from moptipy.api.algorithm import Algorithm 

19from moptipy.api.mo_process import MOProcess 

20from moptipy.api.process import Process 

21 

22 

23class MOAlgorithm(Algorithm): 

24 """A base class for multi-objective optimization algorithms.""" 

25 

26 def solve(self, process: Process) -> None: 

27 """ 

28 Forward to :meth:`solve_mo` and cast `process` to `MOProcess`. 

29 

30 :param process: the process to solve. Must be an instance of 

31 :class:`~moptipy.api.mo_process.MOProcess`. 

32 """ 

33 if not isinstance(process, MOProcess): 

34 raise type_error(process, "process", MOProcess) 

35 self.solve_mo(cast("MOProcess", process)) 

36 

37 def solve_mo(self, process: MOProcess) -> None: 

38 """ 

39 Apply the multi-objective optimization algorithm to the given process. 

40 

41 :param process: the multi-objective process which provides methods to 

42 access the search space, the termination criterion, and a source 

43 of randomness. It also wraps the objective function, remembers the 

44 best-so-far solution, and takes care of creating log files (if 

45 this is wanted). 

46 """