Enhancing the learning capacity of immunological algorithms: a comprehensice study of learning operators