Exact penalties for variational inequalities with applications to nonlinear complementarity problems

In this paper, we present a new reformulation of the KKT system associated to a variational inequality as a semismooth equation. The reformulation is derived from the concept of differentiable exact penalties for nonlinear programming. The best theoretical results are presented for nonlinear complementarity problems, where simple, verifiable, conditions ensure that the penalty is exact. We close the paper with some preliminary computational tests on the use of a semismooth Newton method to solve the equation derived from the new reformulation. We also compare its performance with the Newton method applied to classical reformulations based on the Fischer-Burmeister function and on the minimum. The new reformulation combines the best features of the classical ones, being as easy to solve as the reformulation that uses the Fischer-Burmeister function while requiring as few Newton steps as the one that is based on the minimum.

Métodos de penalidade exacta para resolução de problemas de optimização não linear

In this work we present a classification of some of the existing Penalty Methods (denominated the Exact Penalty Methods) and describe some of its limitations and estimated. With these methods we can solve problems of optimization with continuous, discrete and mixing constrains, without requiring continuity, differentiability or convexity. The boarding consists of transforming the original problem, in a sequence of problems without constrains, derivate of the initial, making possible its resolution for the methods known for this type of problems. Thus, the Penalty Methods can be used as the first step for the resolution of constrained problems for methods typically used in by unconstrained problems. The work finishes discussing a new class of Penalty Methods, for nonlinear optimization, that adjust the penalty parameter dynamically.

El problema del rozamiento en el análisis de estructuras de fábrica mediante modelos de sólidos rígidos

La evaluación de la seguridad de estructuras antiguas de fábrica es un problema abierto.El material es heterogéneo y anisótropo, el estado previo de tensiones difícil de conocer y las condiciones de contorno inciertas. A comienzos de los años 50 se demostró que el análisis límite era aplicable a este tipo de estructuras, considerándose desde entonces como una herramienta adecuada. En los casos en los que no se produce deslizamiento la aplicación de los teoremas del análisis límite estándar constituye una herramienta formidable por su simplicidad y robustez. No es necesario conocer el estado real de tensiones. Basta con encontrar cualquier solución de equilibrio, y que satisfaga las condiciones de límite del material, en la seguridad de que su carga será igual o inferior a la carga real de inicio de colapso. Además esta carga de inicio de colapso es única (teorema de la unicidad) y se puede obtener como el óptimo de uno cualquiera entre un par de programas matemáticos convexos duales. Sin embargo, cuando puedan existir mecanismos de inicio de colapso que impliquen deslizamientos, cualquier solución debe satisfacer tanto las restricciones estáticas como las cinemáticas, así como un tipo especial de restricciones disyuntivas que ligan las anteriores y que pueden plantearse como de complementariedad. En este último caso no está asegurada la existencia de una solución única, por lo que es necesaria la búsqueda de otros métodos para tratar la incertidumbre asociada a su multiplicidad. En los últimos años, la investigación se ha centrado en la búsqueda de un mínimo absoluto por debajo del cual el colapso sea imposible. Este método es fácil de plantear desde el punto de vista matemático, pero intratable computacionalmente, debido a las restricciones de complementariedad 0 y z 0 que no son ni convexas ni suaves. El problema de decisión resultante es de complejidad computacional No determinista Polinomial (NP)- completo y el problema de optimización global NP-difícil. A pesar de ello, obtener una solución (sin garantía de exito) es un problema asequible. La presente tesis propone resolver el problema mediante Programación Lineal Secuencial, aprovechando las especiales características de las restricciones de complementariedad, que escritas en forma bilineal son del tipo y z = 0; y 0; z 0 , y aprovechando que el error de complementariedad (en forma bilineal) es una función de penalización exacta. Pero cuando se trata de encontrar la peor solución, el problema de optimización global equivalente es intratable (NP-difícil). Además, en tanto no se demuestre la existencia de un principio de máximo o mínimo, existe la duda de que el esfuerzo empleado en aproximar este mínimo esté justificado. En el capítulo 5, se propone hallar la distribución de frecuencias del factor de carga, para todas las soluciones de inicio de colapso posibles, sobre un sencillo ejemplo. Para ello, se realiza un muestreo de soluciones mediante el método de Monte Carlo, utilizando como contraste un método exacto de computación de politopos. El objetivo final es plantear hasta que punto está justificada la busqueda del mínimo absoluto y proponer un método alternativo de evaluación de la seguridad basado en probabilidades. Las distribuciones de frecuencias, de los factores de carga correspondientes a las soluciones de inicio de colapso obtenidas para el caso estudiado, muestran que tanto el valor máximo como el mínimo de los factores de carga son muy infrecuentes, y tanto más, cuanto más perfecto y contínuo es el contacto. Los resultados obtenidos confirman el interés de desarrollar nuevos métodos probabilistas. En el capítulo 6, se propone un método de este tipo basado en la obtención de múltiples soluciones, desde puntos de partida aleatorios y calificando los resultados mediante la Estadística de Orden. El propósito es determinar la probabilidad de inicio de colapso para cada solución.El método se aplica (de acuerdo a la reducción de expectativas propuesta por la Optimización Ordinal) para obtener una solución que se encuentre en un porcentaje determinado de las peores. Finalmente, en el capítulo 7, se proponen métodos híbridos, incorporando metaheurísticas, para los casos en que la búsqueda del mínimo global esté justificada. Abstract Safety assessment of the historic masonry structures is an open problem. The material is heterogeneous and anisotropic, the previous state of stress is hard to know and the boundary conditions are uncertain. In the early 50's it was proven that limit analysis was applicable to this kind of structures, being considered a suitable tool since then. In cases where no slip occurs, the application of the standard limit analysis theorems constitutes an excellent tool due to its simplicity and robustness. It is enough find any equilibrium solution which satisfy the limit constraints of the material. As we are certain that this load will be equal to or less than the actual load of the onset of collapse, it is not necessary to know the actual stresses state. Furthermore this load for the onset of collapse is unique (uniqueness theorem), and it can be obtained as the optimal from any of two mathematical convex duals programs However, if the mechanisms of the onset of collapse involve sliding, any solution must satisfy both static and kinematic constraints, and also a special kind of disjunctive constraints linking the previous ones, which can be formulated as complementarity constraints. In the latter case, it is not guaranted the existence of a single solution, so it is necessary to look for other ways to treat the uncertainty associated with its multiplicity. In recent years, research has been focused on finding an absolute minimum below which collapse is impossible. This method is easy to set from a mathematical point of view, but computationally intractable. This is due to the complementarity constraints 0 y z 0 , which are neither convex nor smooth. The computational complexity of the resulting decision problem is "Not-deterministic Polynomialcomplete" (NP-complete), and the corresponding global optimization problem is NP-hard. However, obtaining a solution (success is not guaranteed) is an affordable problem. This thesis proposes solve that problem through Successive Linear Programming: taking advantage of the special characteristics of complementarity constraints, which written in bilinear form are y z = 0; y 0; z 0 ; and taking advantage of the fact that the complementarity error (bilinear form) is an exact penalty function. But when it comes to finding the worst solution, the (equivalent) global optimization problem is intractable (NP-hard). Furthermore, until a minimum or maximum principle is not demonstrated, it is questionable that the effort expended in approximating this minimum is justified. XIV In chapter 5, it is proposed find the frequency distribution of the load factor, for all possible solutions of the onset of collapse, on a simple example. For this purpose, a Monte Carlo sampling of solutions is performed using a contrast method "exact computation of polytopes". The ultimate goal is to determine to which extent the search of the global minimum is justified, and to propose an alternative approach to safety assessment based on probabilities. The frequency distributions for the case study show that both the maximum and the minimum load factors are very infrequent, especially when the contact gets more perfect and more continuous. The results indicates the interest of developing new probabilistic methods. In Chapter 6, is proposed a method based on multiple solutions obtained from random starting points, and qualifying the results through Order Statistics. The purpose is to determine the probability for each solution of the onset of collapse. The method is applied (according to expectations reduction given by the Ordinal Optimization) to obtain a solution that is in a certain percentage of the worst. Finally, in Chapter 7, hybrid methods incorporating metaheuristics are proposed for cases in which the search for the global minimum is justified.

A STUDY OF PENALTY FORMULATIONS USED IN THE NUMERICAL APPROXIMATION OF A RADIALLY SYMMETRIC ELASTICITY PROBLEM

We consider a class of two-dimensional problems in classical linear elasticity for which material overlapping occurs in the absence of singularities. Of course, material overlapping is not physically realistic, and one possible way to prevent it uses a constrained minimization theory. In this theory, a minimization problem consists of minimizing the total potential energy of a linear elastic body subject to the constraint that the deformation field must be locally invertible. Here, we use an interior and an exterior penalty formulation of the minimization problem together with both a standard finite element method and classical nonlinear programming techniques to compute the minimizers. We compare both formulations by solving a plane problem numerically in the context of the constrained minimization theory. The problem has a closed-form solution, which is used to validate the numerical results. This solution is regular everywhere, including the boundary. In particular, we show numerical results which indicate that, for a fixed finite element mesh, the sequences of numerical solutions obtained with both the interior and the exterior penalty formulations converge to the same limit function as the penalization is enforced. This limit function yields an approximate deformation field to the plane problem that is locally invertible at all points in the domain. As the mesh is refined, this field converges to the exact solution of the plane problem.

Augmented Lagrangian method with nonmonotone penalty parameters for constrained optimization

At each outer iteration of standard Augmented Lagrangian methods one tries to solve a box-constrained optimization problem with some prescribed tolerance. In the continuous world, using exact arithmetic, this subproblem is always solvable. Therefore, the possibility of finishing the subproblem resolution without satisfying the theoretical stopping conditions is not contemplated in usual convergence theories. However, in practice, one might not be able to solve the subproblem up to the required precision. This may be due to different reasons. One of them is that the presence of an excessively large penalty parameter could impair the performance of the box-constraint optimization solver. In this paper a practical strategy for decreasing the penalty parameter in situations like the one mentioned above is proposed. More generally, the different decisions that may be taken when, in practice, one is not able to solve the Augmented Lagrangian subproblem will be discussed. As a result, an improved Augmented Lagrangian method is presented, which takes into account numerical difficulties in a satisfactory way, preserving suitable convergence theory. Numerical experiments are presented involving all the CUTEr collection test problems.

Exact Solution of Bayes and Minimax Change-Detection Problems

The challenge of detecting a change in the distribution of data is a sequential decision problem that is relevant to many engineering solutions, including quality control and machine and process monitoring. This dissertation develops techniques for exact solution of change-detection problems with discrete time and discrete observations. Change-detection problems are classified as Bayes or minimax based on the availability of information on the change-time distribution. A Bayes optimal solution uses prior information about the distribution of the change time to minimize the expected cost, whereas a minimax optimal solution minimizes the cost under the worst-case change-time distribution. Both types of problems are addressed. The most important result of the dissertation is the development of a polynomial-time algorithm for the solution of important classes of Markov Bayes change-detection problems. Existing techniques for epsilon-exact solution of partially observable Markov decision processes have complexity exponential in the number of observation symbols. A new algorithm, called constellation induction, exploits the concavity and Lipschitz continuity of the value function, and has complexity polynomial in the number of observation symbols. It is shown that change-detection problems with a geometric change-time distribution and identically- and independently-distributed observations before and after the change are solvable in polynomial time. Also, change-detection problems on hidden Markov models with a fixed number of recurrent states are solvable in polynomial time. A detailed implementation and analysis of the constellation-induction algorithm are provided. Exact solution methods are also established for several types of minimax change-detection problems. Finite-horizon problems with arbitrary observation distributions are modeled as extensive-form games and solved using linear programs. Infinite-horizon problems with linear penalty for detection delay and identically- and independently-distributed observations can be solved in polynomial time via epsilon-optimal parameterization of a cumulative-sum procedure. Finally, the properties of policies for change-detection problems are described and analyzed. Simple classes of formal languages are shown to be sufficient for epsilon-exact solution of change-detection problems, and methods for finding minimally sized policy representations are described.

Study of C- and O-glycosylflavones in sugarcane extracts using liquid chromatography: exact mass measurement mass spectrometry

The flavonoids present in sugarcane (Saccharum officinarum) extracts were analyzed by liquid chromatography - mass spectrometry (LC-MS), and a study of the fragmentation patterns of selected flavonoids was conducted using orthogonal acceleration time-of-flight electrospray ionization mass spectrometry (ESI-oa-ToF MS). Seven C- and O-glycosylflavones were identified in the extracts, namely, schaftoside, isoschaftoside, luteolin-8-C-(rhamnosylglucoside), vitexin, orientin, tricin-7-O-neohesperidoside and tricin-7-O-glucoside. Of these, five were identified in the absence of direct comparison with their respective standards. The described method also permitted the differentiation of the 6-C and 8-C isomeric flavones, schaftoside and isoschaftoside. The combination of fragmentation data and exact mass measurement showed to be complimentary to the HPLC-UV-MS techniques previously utilized for isomers discrimination in sugarcane studies.

Exact time-dependent solutions for a self-regulating gene

The exact time-dependent solution for the stochastic equations governing the behavior of a binary self-regulating gene is presented. Using the generating function technique to rephrase the master equations in terms of partial differential equations, we show that the model is totally integrable and the analytical solutions are the celebrated confluent Heun functions. Self-regulation plays a major role in the control of gene expression, and it is remarkable that such a microscopic model is completely integrable in terms of well-known complex functions.

Evolutionary dynamics on rugged fitness landscapes: Exact dynamics and information theoretical aspects

The parallel mutation-selection evolutionary dynamics, in which mutation and replication are independent events, is solved exactly in the case that the Malthusian fitnesses associated to the genomes are described by the random energy model (REM) and by a ferromagnetic version of the REM. The solution method uses the mapping of the evolutionary dynamics into a quantum Ising chain in a transverse field and the Suzuki-Trotter formalism to calculate the transition probabilities between configurations at different times. We find that in the case of the REM landscape the dynamics can exhibit three distinct regimes: pure diffusion or stasis for short times, depending on the fitness of the initial configuration, and a spin-glass regime for large times. The dynamic transition between these dynamical regimes is marked by discontinuities in the mean-fitness as well as in the overlap with the initial reference sequence. The relaxation to equilibrium is described by an inverse time decay. In the ferromagnetic REM, we find in addition to these three regimes, a ferromagnetic regime where the overlap and the mean-fitness are frozen. In this case, the system relaxes to equilibrium in a finite time. The relevance of our results to information processing aspects of evolution is discussed.

Exact nonequilibrium work generating function for a small classical system

We obtain the exact nonequilibrium work generating function (NEWGF) for a small system consisting of a massive Brownian particle connected to internal and external springs. The external work is provided to the system for a finite-time interval. The Jarzynski equality, obtained in this case directly from the NEWGF, is shown to be valid for the present model, in an exact way regardless of the rate of external work.

A hybrid-mixed finite element formulation for the geometrically exact analysis of three-dimensional framed structures

This paper addresses the development of a hybrid-mixed finite element formulation for the quasi-static geometrically exact analysis of three-dimensional framed structures with linear elastic behavior. The formulation is based on a modified principle of stationary total complementary energy, involving, as independent variables, the generalized vectors of stress-resultants and displacements and, in addition, a set of Lagrange multipliers defined on the element boundaries. The finite element discretization scheme adopted within the framework of the proposed formulation leads to numerical solutions that strongly satisfy the equilibrium differential equations in the elements, as well as the equilibrium boundary conditions. This formulation consists, therefore, in a true equilibrium formulation for large displacements and rotations in space. Furthermore, this formulation is objective, as it ensures invariance of the strain measures under superposed rigid body rotations, and is not affected by the so-called shear-locking phenomenon. Also, the proposed formulation produces numerical solutions which are independent of the path of deformation. To validate and assess the accuracy of the proposed formulation, some benchmark problems are analyzed and their solutions compared with those obtained using the standard two-node displacement/ rotation-based formulation.

Shell curvature as an initial deformation: A geometrically exact finite element approach

An alternative approach for the analysis of arbitrarily curved shells is developed in this paper based on the idea of initial deformations. By `alternative` we mean that neither differential geometry nor the concept of degeneration is invoked here to describe the shell surface. We begin with a flat reference configuration for the shell mid-surface, after which the initial (curved) geometry is mapped as a stress-free deformation from the plane position. The actual motion of the shell takes place only after this initial mapping. In contrast to classical works in the literature, this strategy enables the use of only orthogonal frames within the theory and therefore objects such as Christoffel symbols, the second fundamental form or three-dimensional degenerated solids do not enter the formulation. Furthermore, the issue of physical components of tensors does not appear. Another important aspect (but not exclusive of our scheme) is the possibility to describe exactly the initial geometry. The model is kinematically exact, encompasses finite strains in a totally consistent manner and is here discretized under the light of the finite element method (although implementation via mesh-free techniques is also possible). Assessment is made by means of several numerical simulations. Copyright (C) 2009 John Wiley & Sons, Ltd.

An exact conserving algorithm for nonlinear dynamics with rotational DOFs and general hyperelasticity. Part 2: shells

Following the approach developed for rods in Part 1 of this paper (Pimenta et al. in Comput. Mech. 42:715-732, 2008), this work presents a fully conserving algorithm for the integration of the equations of motion in nonlinear shell dynamics. We begin with a re-parameterization of the rotation field in terms of the so-called Rodrigues rotation vector, allowing for an extremely simple update of the rotational variables within the scheme. The weak form is constructed via non-orthogonal projection, the time-collocation of which ensures exact conservation of momentum and total energy in the absence of external forces. Appealing is the fact that general hyperelastic materials (and not only materials with quadratic potentials) are permitted in a totally consistent way. Spatial discretization is performed using the finite element method and the robust performance of the scheme is demonstrated by means of numerical examples.

An exact conserving algorithm for nonlinear dynamics with rotational DOFs and general hyperelasticity. Part 1: Rods

A fully conserving algorithm is developed in this paper for the integration of the equations of motion in nonlinear rod dynamics. The starting point is a re-parameterization of the rotation field in terms of the so-called Rodrigues rotation vector, which results in an extremely simple update of the rotational variables. The weak form is constructed with a non-orthogonal projection corresponding to the application of the virtual power theorem. Together with an appropriate time-collocation, it ensures exact conservation of momentum and total energy in the absence of external forces. Appealing is the fact that nonlinear hyperelastic materials (and not only materials with quadratic potentials) are permitted without any prejudice on the conservation properties. Spatial discretization is performed via the finite element method and the performance of the scheme is assessed by means of several numerical simulations.

Hybrid and multi-field variational principles for geometrically exact three-dimensional beams

This paper addresses the development of several alternative novel hybrid/multi-field variational formulations of the geometrically exact three-dimensional elastostatic beam boundary-value problem. In the framework of the complementary energy-based formulations, a Legendre transformation is used to introduce the complementary energy density in the variational statements as a function of stresses only. The corresponding variational principles are shown to feature stationarity within the framework of the boundary-value problem. Both weak and linearized weak forms of the principles are presented. The main features of the principles are highlighted, giving special emphasis to their relationships from both theoretical and computational standpoints. (C) 2010 Elsevier Ltd. All rights reserved.