The MoVeS project is a collaborative effort that will explore how stochastic hybrid systems (SHS) methods can be used to address the modeling and control challenges that arise from large-scale, networked systems such as power networks. SHS involve the interaction of continuous dynamics, discrete dynamics and probabilistic uncertainty.

Because of this versatility, SHS are widely recognized as an ideal framework for capturing the intricacies of power networks.

Many case studies have illustrated the potential of SHS in such diverse applications as control of telecommunication networks, air traffic, manufacturing, biology and finance. This has motivated a considerable research effort into the development of modeling analysis and control methods for SHS since the turn of the century, as has given rise several distinct approaches, each with its advantages and disadvantages.

None of the approaches on its own, however, is presently powerful enough to deal with the intricacies and complexity of large-scale, networked power systems. Fortunately, the strengths and weaknesses of the different methods are complementary.

The aim of the MoVeS project is to exploit this complementarity. Our partners bring together a variety of skills and expertise, ranging from formal methods in computer science, to stochastic optimal control, to applications in power networks. 

By bridging the gap and providing formal links between the different classes of methods, we intend to develop novel methodological approaches that are powerful enough to deal with complex, dynamic systems such as power networks.

More specifically:

  1. Modeling and abstraction. We will develop methods, algorithmic solutions and computational tools that support the modeling of large-scale complex systems. Particular emphasis will be placed on the issues of composition and abstraction, which will enable us to model complex systems at different levels of granularity. This makes the systems amenable to different methods that operate at a variety of levels of abstraction, while retaining a clear link between the results.

  2. System analysis and design. We will develop methods, algorithmic solutions and computational tools for analyzing the behaviour of complex systems. Emphasis will be placed on cross-cutting methods. For example, model checking methods will be coupled with tools from the literature on optimal control of stochastic hybrid systems, to extend the former to systems with continuous state spaces and the latter to the analysis of more complex properties encoded by temporal logics.

  3. On-line decision-making and control. We will develop novel methods, algorithmic solutions and computational tools for the control of stochastic hybrid systems. Once again the central theme will be the coupling and cross fertilizations of approaches from different areas.

  4. Power network applications. We will demonstrate the application and potential impact of the novel methods for modeling, analysis and control of hybrid stochastic systems through selected power networks case studies. Power networks provide an ideal testing ground for stochastic hybrid systems, since by nature they involve a tight coupling of continuous dynamics (governing, for example, the power flows and the fluctuations of voltages, frequencies, etc.), discrete dynamics (for example, the positions of switches, protection and isolation devices, transformer taps, etc.), and uncertainty (in the demand for power, but also, with the recent emphasis on renewable energy and distributed generation, in its supply).



VeriSiMPL Toolbox
Verification via biSimulations of Max-Plus-Linear models. This toolbox is used to generate finite abstractions of autonomous and nonautonomous Max-Plus-Linear (MPL) models over R^n. Alessandro Abate & Dieky Adzkiya (TU Delft).

HSCC 2014
April 15-17, 2014
Berlin, Germany
Martin Fraenzle and John Lygeros chair the HSCC Program Committee.