A Feedback Model of Human Personalities

This paper presents a quadrant system and mathematical model for describing the clinical personality disorders, human personality self-images, and human personality behaviours. A proposed biological mechanism that would result in this personality structure means the system can be understood as plausible from first principles. Then similarities of self-images to Q and R in an LQR optimization, and the similarities of behavioural techniques to the gains in a PID feedback control, allow a construction of a computer model. Validity is tested with nonlinear optimization to predict the probability distribution of human behavioural techniques, as described by the Myers-Briggs Temperament Indicator, for different self-images. The results closely match that previously reported from human data. The model directly results in the ability to design control systems with personalities and shows how one would approach programming personalities on robots and chat-bots.