Special Issue Details
With the steady growth in the availability of fast computing machines, control techniques based on algorithmic selection of actions derived from optimisation of a suitable performance index have gained prominence. Model predictive control (or receding horizon control), a framework that is based on such algorithmic selection procedures, has evolved over the years into one of the most useful and applicable control synthesis techniques currently available to a control engineer.
While deterministic and robust versions of model predictive control techniques have pretty much become standardised and are well-documented today, stochastic versions still lack a comprehensive, unified, and systematic treatment. Perhaps a key reason for this lacuna is the fact that the technicalities involved in stochastic model predictive control are significantly more heavy than in the deterministic setting. For instance, while the bare-essential arguments involved in establishing Lyapunov stability of discrete-time deterministic dynamical systems are only a few and are quite classical, the technical arguments and conditions in the theory of stability of stochastic processes are by far more in number, and constitute an active area of research even today. In addition, there exist a variety of notions related to performance and qualitative behaviour that are specific to the stochastic setting, that simply do not exist in the deterministic or the robust cases.
Against this backdrop, this special issue seeks to collect the current state-of-the-art directions in theoretical foundations and applications of stochastic model predictive control, with contributions from the finest researchers who are working in this area