DEbudget: Budget DE

In FilippoPic/DEoptimPIC: DEoptimPIC provides some interesting tools and benchmark functions to test three different versions of the differential evolution algorithm

Description

Differential Evolution algorithm that deals with bound constraints and a budget constraint. Specific for portfolio optimization applications.

Usage

DEbudget(funcname, lo, up, n, NP, n_gen, f, CR, strategy, ...)

Arguments

funcname	function to optimize
lo	lower bound (same for each component of the vector x)
up	upper bound (same for each component of the vector x)
n	problem size
NP	population size (should be ~ four times the size of the problem)
n_gen	number of generations (~ ten times the value of NP)
f	F value (usually from 0.1 to 1.1)
CR	crossover rate (it is a probability, should stay between 0 and 1)
strategy	type of strategy implemented.
...	parameters to pass at objective function to optimize

Value

The output of the function DEbase is a list (of length 3) containing the following elements:

• f_best: the best value found by the algorithm for the last generation

• x_best: the vector which corresponds to the best overall function value

• f: the vector of the n_gen optimal values found by the algorithm at each generation

References

S. Das, S. Mullick, P. N. Suganthan, Recent advances in differential evolution– an updated survey. Swarm and evolutionary computation, vol. 23, 2016, pp. 1–30 Inserire reference per la procedura takkur

Examples

#examplezzzzz

FilippoPic/DEoptimPIC documentation built on Feb. 14, 2022, 5:12 a.m.