A projection-free dynamics for nonsmooth composite optimization
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_ | 1866908611847913472 |
|---|---|
| author | Ni, Wei Qiu, Yangfan Xiao, Yanyan |
| author_facet | Ni, Wei Qiu, Yangfan Xiao, Yanyan |
| contents | This paper proposes a projection-free primal-dual dynamics for the nonsmooth composite optimization problems with equality and inequality constraints. To deal with optimization constraints, this paper departs from the use of gradient projection method, but resorts to the idea of mirror descent to design a continuous-time smooth optimization dynamics which advantageously leads to easier convergence analysis and more efficient numerical simulation. Also, the strategy of proximal augmented Lagrangian (PAL$^†$) is extended to incorporate general convex equality-inequality constraints and the strong convexity-concavity of the primal-dual variables is achieved, ensuring exponential convergence of the resulting algorithm. Furthermore, the convergence result in this paper extends existing exponential convergence which either takes no account of constraints or considers only affine linear constraints, and it also enhances existing asymptotic convergence under convex constraints which unfortunately depends on the complex gradient projection scheme. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_22173 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A projection-free dynamics for nonsmooth composite optimization Ni, Wei Qiu, Yangfan Xiao, Yanyan Optimization and Control This paper proposes a projection-free primal-dual dynamics for the nonsmooth composite optimization problems with equality and inequality constraints. To deal with optimization constraints, this paper departs from the use of gradient projection method, but resorts to the idea of mirror descent to design a continuous-time smooth optimization dynamics which advantageously leads to easier convergence analysis and more efficient numerical simulation. Also, the strategy of proximal augmented Lagrangian (PAL$^†$) is extended to incorporate general convex equality-inequality constraints and the strong convexity-concavity of the primal-dual variables is achieved, ensuring exponential convergence of the resulting algorithm. Furthermore, the convergence result in this paper extends existing exponential convergence which either takes no account of constraints or considers only affine linear constraints, and it also enhances existing asymptotic convergence under convex constraints which unfortunately depends on the complex gradient projection scheme. |
| title | A projection-free dynamics for nonsmooth composite optimization |
| topic | Optimization and Control |
| url | https://arxiv.org/abs/2510.22173 |