Loading...

Networked Cyber-Physical Dynamical Systems with Distributed Optimization-Based Missions

Sarafraz, Mohammad Saeed | 2021

548 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 54404 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Tavazoei, Mohammad Saleh
  7. Abstract:
  8. A new generation of physical and industrial systems is evolving towards communication and information technology as communication and information technologies progress. Cyber-Physical Systems, or CPSs, are physical systems situated in a cyber environment and use communication and information technology to improve efficiency. A CPS is an integrated synthesis of the principles of "control," "communication," and "computation" from an engineering point of view, attempting to provide a comprehensive framework in which all aspects of these three fields are covered. One of the capabilities of CPSs is the ability to deal with and implement a large scale system. A group of control missions in large scale systems can be described in an optimization framework. Multi-agent systems with optimization missions are investigated in this dissertation for this reason.On the other hand, a CPS needs to be optimized as much as possible in accordance with communication and computing practical constraints; This improvement in system performance will reduce energy consumption and increase speed in these systems. In general, performance optimization in a CPS can be considered in three categories: reduction of actuator consumption, communication optimization and computational optimization. In this dissertation, we intend to design methods to reduce communications in multi-agent systems with optimization missions. For this purpose, in this dissertation, two main approaches have been used.The first method is based on reducing communication costs proportional to the weight and number of edges of the communication graph. This approach has been applied to a multi-agent system with uncertain linear dynamics. An integrated algorithm has been proposed to reach a consensus and reduce communication costs between agents.In this dissertation, the second approach for improving efficiency is to use event-based mechanisms. By introducing algorithms, this approach attempts to reduce the number of connections over time. The event-based mechanism for multi-agent systems with a personal objective function is investigated in this dissertation. Specifically, this algorithm is applied to a multi-agent system. Each agent has nonlinear and uncertain dynamics and tries to solve a constraint optimization problem by interacting. In developing classical event-based algorithms that do not require continuous measurement, an algorithm is proposed to control the behaviour of a multi-agent system whose agents have linear and uncertain dynamics. In this system, each agent tries to solve an optimization problem in proportion to the states of other agents and environmental parameters to achieve its target. While implementing the event-based algorithm, we attempted to change the environmental parameters such that the entire network reaches the desired point
  9. Keywords:
  10. Distributed Optimization ; Multiagent System ; Robust Control ; Event-Based Mechanisms ; Communications Optimization ; Cyber-Physical Systems

 Digital Object List

 Bookmark

...see more