Zoom link.

The selection of a subset of nodes is a key strategy for minimizing the error and energy required to steer large-scale complex networks. This is particularly important in real-world systems where only partial control is feasible, and efficient node selection can significantly enhance performance while reducing implementation cost. A fundamental challenge is to identify driver nodes that optimize the network's energy requirements under inherent disturbances, such as those observed in biological systems. However, finding the optimal solution is computationally expensive, if not infeasible. We present an exact solution to this problem, determining the optimal set of driver nodes that simultaneously minimizes the required control energy and steady-state error, while accounting for disturbances. Our findings establish a connection between network controllability, and disturbance rejection by introducing a node-level metric that accounts for local feedback strength as well as structural and global connectivity. We show the superiority of our model on both energy and error measures in simulation on a real biological network for gene therapy applications.

Biography: She is currently a postdoctoral fellow jointly appointed at the University of Toronto and Brock University. Her research focuses on control theory, network dynamics, and the diffusion of innovations in complex systems.