Crate netaheuristics
source · [−]Expand description
Description
This crate’s intended purpose is to offer a modular framework for popular metaheuristics, while simultaneously offering default modules such that the user only needs to focus on the problem-specific aspects of heuristics design.
Custom modules can be built in order to fit the user’s needs.
For now, the following metaheuristics are implemented:
- Variable Neighborhood Search
- Simulated Annealing
- Large Neighborhood Search
Future
The plan for this crate’s future is to assist the user as much as possible in creating metaheuristics. This could mean that other popular metaheuristics are added, or it means that functionality is added to help creating operators.
Modules
Optimization algorithms
Select the next operator to be used
All types required to model termination criteria
Structs
Solution decorated with some metadata
Enums
Evaluation of a proposed candidate
Traits
Evaluate the quality of a solution.
Model of an improvement heuristic based on iterations.
A local search operator returns the neighborhood of its argument.