A5074001836- Advanced Optimization Algorithms Research
- Optimization and Variational Analysis
- Magnetic confinement fusion research
- Sparse and Compressive Sensing Techniques
- Particle accelerators and beam dynamics
- Superconducting Materials and Applications
- Fusion materials and technologies
- Matrix Theory and Algorithms
- Iterative Methods for Nonlinear Equations
- Gyrotron and Vacuum Electronics Research
- Optimization and Mathematical Programming
- Protein Structure and Dynamics
- Nuclear Materials and Properties
- Particle Accelerators and Free-Electron Lasers
- Aerospace Engineering and Control Systems
- Laser-Plasma Interactions and Diagnostics
- Electromagnetic Launch and Propulsion Technology
- Probabilistic and Robust Engineering Design
- Muon and positron interactions and applications
- Stochastic Gradient Optimization Techniques
- Contact Mechanics and Variational Inequalities
- Advanced Multi-Objective Optimization Algorithms
- Nuclear Physics and Applications
- Risk and Portfolio Optimization
- Enzyme Structure and Function
Universidade Estadual de Campinas (UNICAMP)
2015-2025
Universidade Brasil
2016-2024
National Agency for New Technologies, Energy and Sustainable Economic Development
1987-2023
University of Padua
2023
Hospital de Clínicas da Unicamp
2004-2023
Brazilian Society of Computational and Applied Mathematics
2018
Imec the Netherlands
2007
Max Planck Computing and Data Facility
2006
Universidade Estadual Paulista (Unesp)
1998-2003
Max Planck Institute for Plasma Physics
2001-2002
Augmented Lagrangian methods with general lower-level constraints are considered in the present research. These useful when efficient algorithms exist for solving subproblems which only of type. Inexact resolution constrained is considered. Global convergence proved using constant positive linear dependence constraint qualification. Conditions boundedness penalty parameters discussed. The location problems many set nonlinear addressed, employing spectral projected gradient method...
Sequential optimality conditions provide adequate theoretical tools to justify stopping criteria for nonlinear programming solvers. Approximate Karush–Kuhn–Tucker and approximate gradient projection are analysed in this work. These not necessarily equivalent. Implications between different counter-examples will be shown. Algorithmic consequences discussed.
The ITER project requires additional heating by two neutral beam injectors, each accelerating to 1 MV a 40 A of negative deuterium ions, deliver the plasma power about 17 MW for one hour. As these requirements have never been experimentally met, it was recognized as necessary setup test facility, PRIMA (Padova Research on Megavolt Accelerator), in Italy, including full-size ion source, SPIDER, and prototype whole injector, MITICA, aiming develop injectors be installed ITER. This realization...
Every local minimizer of a smooth constrained optimization problem satisfies the sequential approximate Karush--Kuhn--Tucker (AKKT) condition. This optimality condition is used to define stopping criteria many practical nonlinear programming algorithms. It natural ask for conditions on constraints under which AKKT implies KKT. These will be called strict constraint qualifications (SCQs). In this paper we cone-continuity property (CCP) that shown weakest possible SCQ. Its relation other also...
We present two new constraint qualifications (CQs) that are weaker than the recently introduced relaxed constant positive linear dependence (RCPLD) CQ. RCPLD is based on assumption many subsets of gradients active constraints preserve locally. A major open question was to identify exact set whose properties had be preserved locally and would still work as a This done in first CQ, which we call rank subspace component (CRSC) CQ also preserves good RCPLD, such local stability validity an error...
Necessary first-order sequential optimality conditions provide adequate theoretical tools to justify stopping criteria for nonlinear programming solvers. Sequential are satisfied by local minimizers of optimization problems independently the fulfillment constraint qualifications. A new condition this type is introduced in present paper. It proved that a well-established augmented Lagrangian algorithm produces sequences whose limits satisfy condition. Practical consequences discussed.
Many algorithms exist for protein structural alignment, based on internal coordinates or explicit superposition of the structures. These methods are usually successful detecting similarities. However, current practical seldom supported by convergence theories. In particular, although goal each algorithm is to maximize some scoring function, there no method that theoretically guarantees score maximization. A with solid properties would be useful refinement folding maps, and development new...
Sequential optimality conditions for constrained optimization are necessarily satisfied by local minimizers, independently of the fulfillment constraint qualifications. These support employment different stopping criteria practical algorithms. On other hand, when an appropriate property on constraints holds at a point that satisfies sequential condition, such also Karush-Kuhn-Tucker conditions. Those properties will be called strict qualifications in this paper. As consequence, each it is...
In recent years, the theoretical convergence of iterative methods for solving nonlinear constrained optimization problems has been addressed using sequential optimality conditions, which are satisfied by minimizers independently constraint qualifications (CQs). Even though there is a considerable literature devoted to conditions standard optimization, same not true mathematical programs with complementarity constraints (MPCCs). this paper, we show that established suitable analysis...
In the present paper, we prove that augmented Lagrangian method converges to KKT points under quasi-normality constraint qualification, which is associated with external penalty theory. An interesting consequence Lagrange multiplier estimates computed by remain bounded in presence of condition. order establish a more general convergence result, new sequential optimality condition for smooth constrained optimization, called PAKKT, defined. The takes into account sign dual sequence,...
Abstract Necessary optimality conditions for nonlinear programming are discussed in the present research. A new second-order condition is given, which depends on a weak constant rank constraint requirement. We show that practical and publicly available algorithms (www.ime.usp.br/∼egbirgin/tango) of augmented Lagrangian type converge, after slight modifications, to stationary points defined by condition. Keywords: Nonlinear programmingNecessary conditionsConstraint qualificationsPractical...
.The Fritz John (FJ) and Karush–Kuhn–Tucker (KKT) conditions are fundamental tools for characterizing minimizers form the basis of almost all methods constrained optimization. Since seminal works John, Karush, Kuhn, Tucker, FJ/KKT have been enhanced by adding extra necessary conditions. Such an extension was initially proposed Hestenes in 1970s later extensively studied Bertsekas collaborators. In this work, we revisit KKT stationarity standard (smooth) nonlinear programming. We argue that...
Sequential optimality conditions have recently played an important role on the analysis of global convergence optimization algorithms towards first-order stationary points, justifying their stopping criteria. In this article, we introduce a sequential condition that takes into account second-order information and allows us to improve assumptions several algorithms, which is our main goal. We also present companion constraint qualification less stringent than previous associated methods, like...
Minimization of a differentiable function subject to box constraints is proposed as strategy solve the generalized nonlinear complementarity problem (GNCP) defined on polyhedral cone. It not necessary calculate projections that complicate and sometimes even disable implementation algorithms for solving these kinds problems. Theoretical results relate stationary points minimized solutions GNCP are presented. Perturbations also considered, obtained related resolution GNCPs with very general...