Decomposition by Partial Linearization: Parallel Optimization of Multi-Agent Systems
From MaRDI portal
Abstract: We propose a novel decomposition framework for the distributed optimization of general nonconvex sum-utility functions arising naturally in the system design of wireless multiuser interfering systems. Our main contributions are: i) the development of the first class of (inexact) Jacobi best-response algorithms with provable convergence, where all the users simultaneously and iteratively solve a suitably convexified version of the original sum-utility optimization problem; ii) the derivation of a general dynamic pricing mechanism that provides a unified view of existing pricing schemes that are based, instead, on heuristics; and iii) a framework that can be easily particularized to well-known applications, giving rise to very efficient practical (Jacobi or Gauss-Seidel) algorithms that outperform existing adhoc methods proposed for very specific problems. Interestingly, our framework contains as special cases well-known gradient algorithms for nonconvex sum-utility problems, and many blockcoordinate descent schemes for convex functions.
Cited in
(28)- An incremental decomposition method for unconstrained optimization
- Distributed nonconvex constrained optimization over time-varying digraphs
- On the pervasiveness of difference-convexity in optimization and statistics
- Decentralized multi-agent optimization based on a penalty method
- Perturbed proximal primal-dual algorithm for nonconvex nonsmooth optimization
- DC programming and DCA: thirty years of developments
- Hyperparameter estimation for sparse Bayesian learning models
- Asynchronous parallel algorithms for nonconvex optimization
- Variational inequality type formulations of general market equilibrium problems with local information
- Non-cooperative games with minmax objectives
- On the gradient method for solving multi-agent systems
- Computing B-stationary points of nonsmooth DC programs
- Distributed semi-supervised support vector machines
- Decentralized dictionary learning over time-varying digraphs
- Adaptive online distributed optimization in dynamic environments
- Open issues and recent advances in DC programming and DCA
- Inexact zeroth-order nonsmooth and nonconvex stochastic composite optimization and applications
- Convergence rate for diminishing stepsize methods in nonconvex constrained optimization via ghost penalties
- Iteration complexity analysis of block coordinate descent methods
- A block successive upper-bound minimization method of multipliers for linearly constrained convex optimization
- An adaptive partial linearization method for optimization problems on product sets
- Feasible methods for nonconvex nonsmooth problems with applications in green communications
- Ghost penalties in nonconvex constrained optimization: diminishing stepsizes and iteration complexity
- Inexact partial linearization methods for network equilibrium problems
- Diminishing stepsize methods for nonconvex composite problems via ghost penalties: from the general to the convex regular constrained case
- Distributed optimization methods for nonconvex problems with inequality constraints over time-varying networks
- A stochastic successive minimization method for nonsmooth nonconvex optimization with applications to transceiver design in wireless communication networks
- Randomness and permutations in coordinate descent methods
This page was built for publication: Decomposition by Partial Linearization: Parallel Optimization of Multi-Agent Systems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4578978)
