Artificial Immune System Algorithm based on Danger Theory

An artificial immune system algorithm, based on the Danger Theory, is an algorithm which emulates, in a computational context, the defense mechanism of the human immune system when presented with a danger to its wellbeing. Over its evolution the human immune system has developed a complex defense mechanism against entities which are harmful to the human body. This system is divided into two sub-systems that provide different types of defense, the innate immune system that is activated by generic (learned through evolution) harmful entities to the human body, and the adaptive immune system which is activated by specific antigens (molecules which are recognized as dangerous by the immune system). The capacity of the human immune system to flexibly and efficiently respond to threats in the human body is the main inspiration for its artificial intelligence conceptualizations. The study of this system has led researchers to formulate several theories regarding how the immune system works. However, the one we are addressing is the Danger Theory, which was introduced by Matzinger [11, 12] and offered a new perspective to the conceptualization and understanding of the immune system. This model is controversial but it provides answers to several of the problems faced by the previous ones, such as transplant rejections, no immune responses to the body in puberty or spontaneous tumor rejection. The main difference to previous models is that the immune system adapts to a changing self and is activated when it detects danger and not just when in the presence of non-self antigens as previous models proposed. Based on these concepts there are two main ideas behind algorithms using this model: