cma_optim_restarted - functional interface to the CMA-ES optimizer with restarts

Calling Sequence

[xopt, f, out, param] = cma_optim_restarted(costf, x0, sigma0, restarts [, param])

Parameters

• costf see cma_optim()
• x0 the initial solution point, see cma_optim(). When the initial point is chosen randomly, e.g. rand(10,1), consider using a string, i.e. 'rand(10,1)', which is evaluated for each restart anew.
• sigma0 see cma_optim()
• restarts maximal number of restarts.
• param see cma_optim()
• xopt see cma_optim()
• f see cma_optim()
• out see cma_optim()
• param see cma_optim()

Description

Calls the function cma_optim several times with increasing population size parameter opt.lambda. For each restart the population size lambda is increased by a factor of two. No further restarts are conducted when the target fitness or the maximum number of function evaluation maxfunevals was reached.

Examples

see cma_optim

See Also

cma_optim ,   cma_new ,   cma_plot ,   optim ,  

Authors

Nikolaus Hansen

Bibliography

Used Function