Optimization approaches for network resource management

#network-resource-management #digital-twin #scheduling #synchronization
Share

Optimization problems in network traffic scheduling and resource management are often challenging due to uncertainty, nonconvexity, large decision spaces, and other system-specific constraints. In this talk, I will present several of our recently developed approaches for addressing these challenges.

The first problem considers computation offloading over time-varying wireless channels and seeks the optimal offloading start time while satisfying task completion deadlines. A time-dilated absorbing Markov model is developed, and optimal stopping theory for Markov decision processes is used to derive the optimal offloading policy.

The second problem investigates the placement of digital twins on edge servers. The objective is to minimize the synchronization time between physical systems (PSs) and their digital twins (DTs), subject to the communication and computation requirements. The problem is formulated as a multi-commodity quickest flow evacuation problem, followed by an unsplittable-flow rounding procedure.

The third problem considers privacy-preserving channel resource sharing using federated optimization and digital twins. A mixed-integer continuous optimization problem is formulated and relaxed into a continuous problem. An approximate solution is then obtained through dependent rounding followed by convex re-optimization.



  Date and Time

  Location

  Hosts

  Registration



  • Add_To_Calendar_icon Add Event to Calendar

Loading virtual attendance info...

  • Contact Event Hosts


  Speakers

Dongmei of McMaster University

Topic:

Optimization approaches for network resource management

Optimization problems in network traffic scheduling and resource management are often challenging due to uncertainty, nonconvexity, large decision spaces, and other system-specific constraints. In this talk, I will present several of our recently developed approaches for addressing these challenges.

The first problem considers computation offloading over time-varying wireless channels and seeks the optimal offloading start time while satisfying task completion deadlines. A time-dilated absorbing Markov model is developed, and optimal stopping theory for Markov decision processes is used to derive the optimal offloading policy.

The second problem investigates the placement of digital twins on edge servers. The objective is to minimize the synchronization time between physical systems (PSs) and their digital twins (DTs), subject to the communication and computation requirements. The problem is formulated as a multi-commodity quickest flow evacuation problem, followed by an unsplittable-flow rounding procedure.

The third problem considers privacy-preserving channel resource sharing using federated optimization and digital twins. A mixed-integer continuous optimization problem is formulated and relaxed into a continuous problem. An approximate solution is then obtained through dependent rounding followed by convex re-optimization.

Biography:

Dongmei Zhao received the Ph.D degree in Electrical and Computer Engineering from University of Waterloo, Waterloo, Ontario, Canada in June 2002. In July 2002 she joined the Department of Electrical and Computer Engineering at McMaster University, where she is a professor. She is currently an Associate Editor for the IEEE Internet of Things Journal and a Distinguished Lecturer of the IEEE Vehicular Technology Society. Her recent research interests focus on autonomous networking and digital twins.

Email:

Address:1280 Main St W, , Hamilton, Ontario, Canada, L8S 1C7