It would be great to extend the current implementation of the Swarm::globalBestParticle to some kind of fixed-capacity (e.g. https://stackoverflow.com/questions/56334492/c-create-fixed-size-queue) so that a more complex normal stop criterion could be implemented. The criterion might decide upon more than just 2 best values, e.g. deviance from the average for last 10 best values.
It would be great to extend the current implementation of the
Swarm::globalBestParticleto some kind of fixed-capacity (e.g. https://stackoverflow.com/questions/56334492/c-create-fixed-size-queue) so that a more complex normal stop criterion could be implemented. The criterion might decide upon more than just 2 best values, e.g. deviance from the average for last 10 best values.https://github.com/egi-megi/rojCzastek_Szczepanski_Jurkiewicz_Pikulinski/blob/c947887807dd9a998d25dd3e682aaaa6f7fe5c78/include/Swarm.h#L28