Separable Approximations of Optimal Value Functions and Their Representation by Neural Networks
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914003426476032 |
|---|---|
| author | Sperl, Mario Saluzzi, Luca Kalise, Dante Grüne, Lars |
| author_facet | Sperl, Mario Saluzzi, Luca Kalise, Dante Grüne, Lars |
| contents | The use of separable approximations is proposed to mitigate the curse of dimensionality related to the approximation of high-dimensional value functions in optimal control. The separable approximation exploits intrinsic decaying sensitivity properties of the system, where the influence of a state variable on another diminishes as their spatial, temporal, or graph-based distance grows. This property allows the efficient representation of global functions as a sum of localized contributions. A theoretical framework for constructing separable approximations in the context of optimal control is proposed by leveraging decaying sensitivity in both discrete and continuous time. Results extend prior work on decay properties of solutions to Lyapunov and Riccati equations, offering new insights into polynomial and exponential decay regimes. Connections to neural networks are explored, demonstrating how separable structures enable scalable representations of high-dimensional value functions while preserving computational efficiency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_08559 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Separable Approximations of Optimal Value Functions and Their Representation by Neural Networks Sperl, Mario Saluzzi, Luca Kalise, Dante Grüne, Lars Optimization and Control 49L20, 68T07, 93C41 The use of separable approximations is proposed to mitigate the curse of dimensionality related to the approximation of high-dimensional value functions in optimal control. The separable approximation exploits intrinsic decaying sensitivity properties of the system, where the influence of a state variable on another diminishes as their spatial, temporal, or graph-based distance grows. This property allows the efficient representation of global functions as a sum of localized contributions. A theoretical framework for constructing separable approximations in the context of optimal control is proposed by leveraging decaying sensitivity in both discrete and continuous time. Results extend prior work on decay properties of solutions to Lyapunov and Riccati equations, offering new insights into polynomial and exponential decay regimes. Connections to neural networks are explored, demonstrating how separable structures enable scalable representations of high-dimensional value functions while preserving computational efficiency. |
| title | Separable Approximations of Optimal Value Functions and Their Representation by Neural Networks |
| topic | Optimization and Control 49L20, 68T07, 93C41 |
| url | https://arxiv.org/abs/2502.08559 |