Title: Gradient-based stochastic optimization under chance constraints

Speaker: Felisa Vazquez-Abad
City University of New York

Summary:
Optimization under probability (or chance) constraints is known to be a challenging problem. Generalizing our preliminary results for one-dimensional problems, we study here the general model where the constraints may depend on a stochastic process parametrized by the control variable of the optimization problem. The distribution of the underlying process is not known and only samples of it are available, so the optimization algorithm must be built for streaming data.

Instead of formulating the constraints as probability constraints, we use the alternative form in terms of quantile constraints, which is known to yield more efficient gradient-based algorithms. Common methods for solving such problems make use of advanced statistical functional estimation of the distribution function directly, using approximations that are provably convergent.

Instead, we propose to build statistical estimators for both the quantile and its derivative, in order to drive the optimization procedure. We discuss the computational challenges: because quantiles are commonly estimated using the complete history of samples obtained, they require an undesirable super-linear running time and unbounded memory. We discuss various approaches to overcome this difficulty.

Biography:
Felisa Vázquez-Abad is Professor of Computer Science at City University New York (CUNY). She is interested in stochastic  optimization and computer simulation of complex systems under uncertainty, primarily to build efficient self-regulated learning systems. She has applied novel techniques for simulation and optimization in telecommunications, transportation, medical and biological models, finance and insurance. She is interested by real life problems. She co-authored a US patent for an optical network switch and was research consultant to the Melbourne Airport. She has published more than 100 research papers and her textbook on “Optimization and Learning via Stochastic Gradient Search” is being published by Princeton University Press in 2025.

She obtained a B.Sc. in Physics (1983) and a M.Sc. in Statistics and Operations Research (1984) from UNAM (Mexico). In 1989 she obtained a Ph.D. in Applied Mathematics from Brown University. She was postdoctoral researcher at the INRS-Telecom in Montreal, Canada from 1990 to 1993. She was a professor at the University of Montreal (1993-2004), and then a professor at the University of Melbourne, until 2009 that she moved to New York to join the CUNY Faculty.

In 2000, she was a recipient of the Jacob Wolfowitz award for advances in the mathematical and management sciences.Her team was a finalist in the MEDSTART competition (2014) and she lead the award-winning Hunter Hawks team in the CUNY-IBM Watson Competition (2017). She has participated in Grant Selection Committees and has been Associate Editor for IEEE Transactions on Automatic Control, Management Science, and Operations Research Letters, Area Editor of the ACM Transactions on Computer Modeling and Simulation, and web editor of the INFORMS College on Simulation. She actively participates in events and programs to encourage women and minority students to succeed in Computer Science.

—

This event is Hybrid: Seminar to start 16:00 on Zoom and in-person
Join us in-person before the seminar begins for a meet and greet at: 15:30 
** As space is limited, it is essential to register for in-person attendance at: https://go.unimelb.edu.au/t5np
Location for in person: Melbourne Connect, Level 8, Room 8109, 700 Swanston Street, Carlton 3053

SEMINAR: WED 5 MARCH 2025 15:30 – 17:00 (AEDT, Melbourne Time)
ZOOM MEETING ID: 873 1557 5255; PASSWORD: 778635