Discrete-time dynamic modeling for software and services composition as an extension of the Markov chain approach

Abstract

Discrete Time Markov Chains (DTMCs) and Continuous Time Markov Chains (CTMCs) are often used to model various types of phenomena, such as, for example, the behavior of software products. In that case, Markov chains are widely used to describe possible time-varying behavior of “self-adaptive” software systems, where the transition from one state to another represents alternative choices at the software code level, taken according to a certain probability distribution. From a control-theoretical standpoint, some of these probabilities can be interpreted as control signals and others can just be observed. However, the translation between a DTMC or CTMC model and a corresponding first principle model, that can be used to design a control system is not immediate. This paper investigates a possible solution for translating a CTMC model into a dynamic system, with focus on the control of computing systems components. Notice that DTMC models can be translated as well, providing additional information.