| RUS ENG |
| Полная версия |
|
|
|
| СЕМИНАРЫ |
|
Seminar on Analysis, Differential Equations and Mathematical Physics
|
|||
|
|
|||
|
Intelligent Self-Adaptive Methods for Stochastic Optimization A. Sifaleras |
|||
|
Аннотация: Stochastic optimization methods, such as differential evolution and genetic algorithms, provide powerful approaches for solving computationally hard problems in areas ranging from combinatorial optimization to numerical approximation. A fundamental challenge in these methods is parameter selection: performance is highly sensitive to parameter choices, traditionally requiring extensive offline tuning experiments. This talk presents recent developments in self-adaptive optimization methods, where parameters adjust automatically during the stochastic search process. We examine the principles behind these adaptive mechanisms using the variable neighborhood search metaheuristic and discuss open research problems in this area. Язык доклада: английский Website: https://msrn.tilda.ws/sl |
|||