malschains {Rmalschains}

Perform optimization with the MA-LS-Chains algorithm

Description

This is the main function of the package. It minimizes the output of the function fn (for maximization, change the sign of the output of fn).

Usage

malschains(
  fn,
  lower,
  upper,
  dim,
  maxEvals = 10 * control$istep,
  verbosity = 2,
  initialpop = NULL,
  control = malschains.control(),
  seed = NULL,
  env
)

Arguments

fn

The function to minimize.

lower

The lower bound (or bounds) of the search domain.

upper

The upper bound (or bounds) of the search domain.

dim

The dimension of the problem (if lower and upper are vectors it is not needed).

maxEvals

The maximal number of evaluations of the fitness function.

verbosity

Set the verbosity level. Currently, meaningful values are 0, 1, 2

initialpop

An initial population for the evolutionary algorithm can be submitted (as a matrix). Here, prior knowledge can be introduced to get better results from the algorithm.

control

A list containing the main options of the algorithm. See malschains.control.

seed

A seed value for the random number generator.

env

The environment in which to evaluate the fitness function. If not given, it is generated.

Details

The output of the function when run with verbosity=2 is the following:

• EA::PopFitness The fitness of the best, the one at the 1st quartile, the one at the 3rd quartile, and the worst individual.

• EA::Improvement Improvement of the individuals at the according ranked positions in the population (best, 1st quartile, 3rd quartile, worst).

• LS The number of the individual which is improved on (in braces), its fitness before and after application of the LS procedure, and their difference.

• EABest If the best fitness present in the population changed: same as LS.

Value

the function returns a list containing the best individual, sol, and its fitness. Furthermore, it contains some information on the optimization process, which can be seen using print.malschains.

References

Molina, D., Lozano, M., Sánchez, A.M., Herrera, F. Memetic algorithms based on local search chains for large scale continuous optimisation problems: MA-SSW-Chains (2011) Soft Computing, 15 (11), pp. 2201-2220.

Molina, D., Lozano, M., García-Martínez, C., Herrera, F. Memetic algorithms for continuous optimisation based on local search chains (2010) Evolutionary Computation, 18 (1), pp. 27-63.