Optimisation is the cornerstone of many engineering systems and cyber-physical systems including smart homes, energy grids, and intelligent transportation systems. In many situations however, state-of-the-art optimisation algorithms may fail to provide acceptable (and feasible) solutions e.g. because of the scale of the problem, because the problem is continuously changing in time, or because the problem is ill-posed (i.e., depends on a cost function that is unknown).
This talk will focus on how to build an online algorithm to solve a time-varying optimization problem with an objective that comprises a known time-varying cost and an unknown function. This problem structure arises in cyber-physical and social systems where the known function captures time-varying engineering costs, and the unknown function models user’s satisfaction; in this context, the objective is to strike a balance between given performance metrics and user’s satisfaction that has to be learn online and concurrently with the execution of the optimisation algorithm. The applications in this area stemming from smart energy grids and vehicle control will be described.
The talk main references are the papers:
About the Speaker
Dr Andrea Simonetto is a research staff member in the AI & Quantum team of IBM Research Ireland, in Dublin. He received his PhD in systems and control from Delft University of Technology, The Netherlands in 2012, and spent 3+1 years as as postdoc, first in the signal processing group in the electrical engineering department in Delft, then in the applied mathematics department of the Université catholique de Louvain, in Belgium.
He was a visiting researcher at Carnegie Mellon University, University of Pennsylvania, and KTH, Sweden and joined IBM Research in February 2017.
His interests span optimisation, control, and signal processing, with applications in smart energy, smart transportation, personalised health, and quantum computing.