setting initial parameters for scipy.differential_evolution

Question

In a differential evolution optimization algorithm from scipy, is it possible to set initial parameters if I have some good starting points?

For example, suppose i know the best x is [0.1, 0.5,0.3], is it possible to add it inside scipy.differential_evolution

Answer 1

The scipy.optimize.differential_evolution function has two parameters you can work with:

1. bounds : sequence

   Bounds for variables. (min, max) pairs for each element in x, defining the lower and upper bounds for the optimizing argument of func. [CUT]

   You could try:

   >>> bounds = [(0.0, 0.2), (0.4, 0.6), (0.2, 0.4)]
   >>> result = differential_evolution(my_func, bounds)
   

2. Recent versions of the library (from SciPy v1.1.0) allow to specify which type of population initialization is performed via the init keyword:

   init : str or array-like, optional

   Specify which type of population initialization is performed. Should be one of:

   • ‘latinhypercube’
   • ‘random’
   • array specifying the initial population. The array should have shape (M, len(x)), where len(x) is the number of parameters. init is clipped to bounds before use.

   [CUT]

   Use of an array to specify a population subset could be used, for example, to create a tight bunch of initial guesses in an location where the solution is known to exist, thereby reducing time for convergence.