New models for immune mechanism diagnosis

Abstract

In this paper we introduce a discrete event system model of immune systems of mammalians using Markov Decision Processes based on Petri Nets Models. Based on immunological principles we propose an approach in order to study the mechanisms that govern the immune system's functionality. A two-module algorithm that launches a specific action against an anomalous situation is developed. The Petri nets tools are assumed in this approach. Also, Markov Decision Processes (MDPs) with a truncated state space to the problem with infinite state space is considered. We also propose a new algorithm to build a large model (e.g., a macro-model) of immune mechanisms of mammalians. We show that an optimal stationary policy exists and we apply the results of [1] to a dynamic scheduling problem of the immunological response to external stimuli. © 2008 Springer-Verlag.