Description

WAGGON focuses on optimisation methods, and Optimiser is a base class for optimisation algorithms. It contains common methods and properties, e.g., optimise that runs the optimisation loop until the chosen error, error_type, is small enough, opt_eps, and create_candidates that samples candidate points using Latin Hyperube sampling.

The class can be inherited for implementing specific approaches. Currently waggon.optim contains surrogate-based optimisation, SurrogateOptimiser (described in the following section). New methods will be added as our research continues and can be suggested via a pull request.

Usage

from waggon.optim import Optimiser