Online Seminar by Prof. Paulo J.S Silva
- 【Date & Time】
- August 29th, 10:00 AM
- 【Place】
- Zoom
- Link: https://us06web.zoom.us/j/84253627171?pwd=QTYyZW4vaHhGK0tOaisrNm15aXJPQT09
Meeting ID: 842 5362 7171
Passcode: 642187 - 【Speaker】
- Paulo J.S Silva (State University of Campinas)
- 【Title】
- Sequential Optimality Conditions and Optimization Algorithm Termination: The Case of the Scaled Stopping Criterion
- 【Abstract】
- Modern continuous optimization leans heavily on optimality conditions, notably the renowned KKT conditions. Initially introduced by Karush, these conditions were later rediscovered and widely popularized by Kuhn and Tucker (KKT). They serve a dual purpose: on one hand, they define good candidates for a solution, and on the other, they act as foundational elements in the developing optimization algorithms. Recognizing the inherent approximation nature of sequences produced by algorithms leads naturally to sequential counterparts. This talk will delve into this evolution, presenting its implications for understanding the convergence of computational methods and their respective stopping criteria. A key discussion point will be an analysis of a scaled version of KKT, frequently employed in continuous optimization routines.
