Design admissibility and de la Garza phenomenon in multifactor experiments