Global ETD Search

Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.

Results

Showing 1 to 20 of 20 for “"Subgradient methods"”.

  1. Inexact subgradient methods.

    … evaluation of the objective function and its subgradients can be computationally burdensome. For example, in a stochastic program, the objective function is typically defined through a multi-dimensional integration. Solution procedures for stochastic programs are usually based on functional …

    arizona-thes Repository record for Inexact subgradient methods. (opens in a new tab)

  2. SOME PRIMAL-DUAL THEORY FOR SUBGRADIENT METHODS FOR STRONGLY CONVEX OPTIMIZATION

    We consider (stochastic) subgradient methods for strongly convex but potentially nonsmooth non-Lipschitz optimization. We provide new equivalent dual descriptions (in the style of dual averaging) for the classic subgradient method, the proximal subgradient method, and the switching subgradient …

    jhu Repository record for SOME PRIMAL-DUAL THEORY FOR SUBGRADIENT METHODS FOR STRONGLY CONVEX OPTIMIZATION (opens in a new tab)

  3. Bregman generalized subgradient projected algorithms and applications

    Projected subgradient methods were thought as ideal algorithms for constrained minimization problems. But they are often hampered by the computational complexity of the projection operator. The generalized Bregman distance has been identified as a potential solution to this issue. This thesis …

    ubc Repository record for Bregman generalized subgradient projected algorithms and applications (opens in a new tab) Download PDF for Bregman generalized subgradient projected algorithms and applications (opens in a new tab)

  4. Saddle Avoidance, Asymptotic Normality, and Exponential Acceleration in Nonsmooth Optimization

    … are the foundation for empirically successful methods in modern fields, such as artificial intelligence and data science. Although classical optimization theory provides guarantees for functions with smoothness or convexity, a significant portion of modern problems do not possess any of these. …

    cornell Repository record for Saddle Avoidance, Asymptotic Normality, and Exponential Acceleration in Nonsmooth Optimization (opens in a new tab)

  5. Geometry and algorithms for signal recovery: from convex duality to non-convex formulations

    … Second, we investigate a generalization of subgradient methods, originally designed for convex functions, to functions that are only weakly convex but enjoy the geometric advantage of sharpness. We see that subgradient methods on this class of functions converge at a local linear rate. …

    washington Repository record for Geometry and algorithms for signal recovery: from convex duality to non-convex formulations (opens in a new tab)

  6. Cost minimization in multi−commodity multi−mode generalized networks with time windows

    … bounds based on Lagrangian relaxations with subgradient methods. Phase 2 applies two methods, early due date with overdue-date costs and total transportation costs, to search for an improved upper bound. Two application networks are used to test HA for small and medium-scale problems. A …

    tamu Repository record for Cost minimization in multi−commodity multi−mode generalized networks with time windows (opens in a new tab)

  7. FIRST-ORDER METHODS FOR NONSMOOTH NONCONVEX FUNCTIONAL CONSTRAINED OPTIMIZATION WITH OR WITHOUT SLATER POINTS

    … similar to those of "Quadratically regularized subgradient methods for weakly convex optimization with weakly convex constraints" by Ma et al. and "Stochastic first-order methods for convex and nonconvex functional constrained optimization" by Boob et al. (whose guarantees both assume …

    jhu Repository record for FIRST-ORDER METHODS FOR NONSMOOTH NONCONVEX FUNCTIONAL CONSTRAINED OPTIMIZATION WITH OR WITHOUT SLATER POINTS (opens in a new tab)

  8. Regret bounds without Lipschitz continuity : online learning with relative-Lipschitz losses

    … classical counterparts. It has been shown that subgradient methods in the relative setting have performance analogous to their performance in the classical setting. In this work, we consider OCO for relative Lipschitz and relative strongly con- vex functions. We extend the known regret bounds …

    ubc Repository record for Regret bounds without Lipschitz continuity : online learning with relative-Lipschitz losses (opens in a new tab) Download PDF for Regret bounds without Lipschitz continuity : online learning with relative-Lipschitz losses (opens in a new tab)

  9. Algorithms for Online Marketplaces: New Approaches to Order Fulfillment and Recommendation Systems

    … strategies, prophet inequalities, and subgradient methods. Our approaches not only achieve asymptotic optimality and strong approximation guarantees in the multi-item fulfillment setting, but also provide insights on how to construct robust policies in scenarios where you have limited …

    duke Repository record for Algorithms for Online Marketplaces: New Approaches to Order Fulfillment and Recommendation Systems (opens in a new tab)

  10. Quantum Advantage via Physics-Inspired Methods in Optimization: Empirical and Theoretical Analysis

    … the precise conditions under which these methods provide a provable advantage over state-of-the-art classical algorithms are still largely unexplored. In this thesis, we make progress toward answering this question. First, we establish the global convergence of QHD for nonsmooth continuous …

    maryland Repository record for Quantum Advantage via Physics-Inspired Methods in Optimization: Empirical and Theoretical Analysis (opens in a new tab)

  11. Wireless communications in low-earth orbit (LEO) satellite systems

    … it into simpler subproblems solvable via subgradient methods. Compared to conventional strict protection constraints—which lead to severe under-utilization of the secondary system—our proposed interference protection approach provides much-needed flexibility without significantly …

    texas Repository record for Wireless communications in low-earth orbit (LEO) satellite systems (opens in a new tab)

  12. Design and operation of electricity markets: dynamics, uncertainty, pricing and competition

    … of the algorithm compares well with standard subgradient methods on the examples considered. Finally, the convex hull pricing scheme is compared with the currently employed marginal-cost pricing scheme in duopolistic power markets in which firms submit their offer functions while abiding by …

    uiuc Repository record for Design and operation of electricity markets: dynamics, uncertainty, pricing and competition (opens in a new tab)

  13. Offset-Based Beamforming: A New Approach to Robust Downlink Transmission

    … provides near-optimal power loading. Using subgradient methods, variants of the offset maximization algorithms that can accommodate per-antenna power constraints (PAPCs) are developed. Furthermore, the resulting offset-based power loading method can be combined with the maximum ratio …

    mcmaster Repository record for Offset-Based Beamforming: A New Approach to Robust Downlink Transmission (opens in a new tab)

  14. Combatting ill-conditioning and heavy-tailed noise in non-convex optimization

    … challenge for optimization with first-order methods. This is especially the case for stochastic optimization, where the classical preconditioners - designed to compensate ill-conditioned curvatures - are hard to construct due to the inherent variability of stochastic gradients. Compounding …

    cornell Repository record for Combatting ill-conditioning and heavy-tailed noise in non-convex optimization (opens in a new tab)

  15. Distributed online algorithms for energy management in smart grids

    … for distributed economic dispatch based on Subgradient method and Alternating Direction Method of Multipliers (ADMM), both designed to be agnostic with any initialization vector. The proposed distributed online solutions leverage a dynamic average consensus algorithm to track the …

    utc Repository record for Distributed online algorithms for energy management in smart grids (opens in a new tab)

  16. Accelerated first-order optimization methods using inertia and error bounds

    … applications. The focus is on first-order methods which have low per-iteration complexity and can exploit problem structure to a high degree. First-order methods have the capacity to address large-scale problems for which all alternative methods fail. However, first-order methods can take …

    uiuc Repository record for Accelerated first-order optimization methods using inertia and error bounds (opens in a new tab)

  17. Nonconvex Trajectory Optimization using Trajectory Sensitivities: Application to Personalized Autonomous Driving

    … this goal. However, the performance of these methods is highly dependent on sample efficiency, and it is usually difficult to enforce safety guarantees. To mitigate these difficulties, this thesis proposes an autonomous vehicle control framework in the form of a parameterized nonconvex …

    waterloo Repository record for Nonconvex Trajectory Optimization using Trajectory Sensitivities: Application to Personalized Autonomous Driving (opens in a new tab)

  18. Tight Flow-Based Formulations for the Asymmetric Traveling Salesman Problem and Their Applications to some Scheduling Problems

    … to (near-) optimality by using deflected subgradient methods on Lagrangian dual formulations. We solve the LP relaxation of our tightest formulation, ATSP6, to (near-) optimality by using a deflected subgradient algorithm with average direction strategy (SA_ADS) (see Sherali and Ulular …

    vt Repository record for Tight Flow-Based Formulations for the Asymmetric Traveling Salesman Problem and Their Applications to some Scheduling Problems (opens in a new tab)

  19. Nondifferentiable Optimization of Lagrangian Dual Formulations for Linear Programs with Recovery of Primal Solutions

    … which simplex as well as interior point based methods can tend to be ineffective. In contrast, Lagrangian relaxation or dual formulations, when applied in concert with suitable primal recovery strategies, have the potential for providing quick bounds as well as enabling useful branching …

    vt Repository record for Nondifferentiable Optimization of Lagrangian Dual Formulations for Linear Programs with Recovery of Primal Solutions (opens in a new tab)

  20. Discrete Two-Stage Stochastic Mixed-Integer Programs with Applications to Airline Fleet Assignment and Workforce Planning Problems

    … non-smooth Lagrangian dual problems using subgradient methods in the bounding process, which turns out to be computationally very expensive. We begin with proposing a decomposition-based branch-and-bound (DBAB) algorithm for solving two-stage stochastic programs having 0-1 mixed-integer …

    vt Repository record for Discrete Two-Stage Stochastic Mixed-Integer Programs with Applications to Airline Fleet Assignment and Workforce Planning Problems (opens in a new tab)