Introducing

Real-Time Physical Systems


RTPS


A research about real-time scheduling of physical systems

Video introduction: RTPS and Electric Load Scheduling



What a RTPS is

A Real-Time Physical System is about scheduling physical systems as they were computing tasks. A real-time scheduler is used as controller of the system. Traditional scheduling algorithms, such as the Earliest Deadline First (EDF), can be used by the scheduler. Physical systems are modeled as dynamical systems, having the real-time schedule as input signal.


Main idea: electric power - computing power analogy

The scheduler inside the Operating System makes many programs run "simultaneously" on our computers. In the same way, in our houses many electric devices are active at the same time. Applying the super-efficient scheduling techniques of our computers to our houses can be very useful!


RTPS benefits

  • It is an efficient way to find an approximated solution to the peak load minimization problem
  • It allows to achieve a peak load close to the theoretical minimum
  • Scalability: it has a low computational cost, so it can manage hundreds of loads
  • It can dynamically manage changes in the load set
  • It can leverage on 40+ years of research in the field of real-time scheduling algorithms

RTPS model in a nutshell

A RTPS is composed by:

  • a set of ON/OFF loads that consume power
  • physical quantities of interest (eg. room temperature, air pressure, ...)
  • user requirements, i.e. bounds on quantities of interest
  • the RT scheduler, which generates the switching signal
  • a switched hybrid system, which describes the physical system dynamics
Evolution over time example of a RTPS composed by one load.
                         The state variable decreases while the switching signal is equal to 1 and it increases when
                         the switching signal is equal to 0.

RTPS design

For many dynamical systems, there are theoretical results showing how to choose the timing parameters for designing the scheduling algorithm such that the evolution of the physical system is under control.


RTPS example

Comparison of the total power consumption of 100 air conditioners controlled by an hysteresis controller and by a RTPS controller. It is remarkable the flattening effect due to the coordinated scheduling. Flattern power demand has a number of advantages, including a peak load reduction.