freegs.optimise.optimise

Contents

freegs.optimise.optimise#

freegs.optimise.optimise(eq, controls, measure, maxgen=10, N=10, CR=0.3, F=1.0, monitor=None)[source]#

Use Differential Evolution to optimise an Equilibrium https://en.wikipedia.org/wiki/Differential_evolution

Parameters:

  • eq – Equilibrium to be used as starting solution. These are deep copied, and passed as arguments to controls and measure functions

  • controls –

    List of control objects. These objects must have methods:

    • .set(eq, value) which modifies the given Equilibrium

    • .get(eq) which returns the value from the Equilibrium

  • measure(eq) – Function which returns a score (value) for a given Equilibrium. The optimiser tries to minimise this value.

  • maxgen – Maximum number of generations

  • N – Population size (must be >= 4)

  • CR – Crossover probability (must be in [0,1])

  • F – Differential weight (must be in [0,2])

  • monitor(generation –

    A function to be called each generation with the best Equilibrium and the whole population

    • generation = integer

    • best = (score, object)

    • population = [(score, object)]

  • best –

    A function to be called each generation with the best Equilibrium and the whole population

    • generation = integer

    • best = (score, object)

    • population = [(score, object)]

  • population) –

    A function to be called each generation with the best Equilibrium and the whole population

    • generation = integer

    • best = (score, object)

    • population = [(score, object)]

Returns:

The Equilibrium with the lowest measure (score).

Return type:

Equilibrium

Contents