An articulated figure is often modeled as a set of rigid segments connected with joints. Its configuration can be altered by varying the joint angles. Although it is straightforward to compute figure configurations given joint angles (forward kinematics), it is not so to find the joint angles for a desired configuration (inverse kinematics). Since the inverse kinematics problem is of special importance to an animator wishing to set a figure to a posture satisfying a set of positioning constraints, researchers have proposed many approaches. But when we try to follow these approaches in an interactive animation system where the object to operate on is as highly articulated as a realistic human figure, they fail in either generality or performance, and so a new approach is fostered. Our approach is based on nonlinear programming techniques. It has been used for several years in the spatial constraint system in the Jack TM human figure simulation software developed at the Compute...

. In this note we show how the implicit filtering algorithm can be coupled with the BFGS quasi-Newton update to obtain a superlinearly convergent iteration if the noise in the objective function decays sufficiently rapidly as the optimal point is approached. We show how known theory for the noise-free case can be extended and thereby provide a partial explanation for the good performance of quasi-Newton methods when coupled with implicit filtering. Key words. noisy optimization, implicit filtering, BFGS algorithm, superlinear convergence AMS subject classifications. 65K05, 65K10, 90C30 1. Introduction. In this paper we examine the local and global convergence behavior of the combination of the BFGS [4], [20], [17], [23] quasi-Newton method with the implicit filtering algorithm. The resulting method is intended to minimize smooth functions that are perturbed with low-amplitude noise. Our results, which extend those of [5], [15], and [6], show that if the amplitude of the noise decays ...

Mixture of experts (ME) is a modular neural network architecture for supervised learning. A double-loop Expectation-Maximization (EM) algorithm has been introduced to the ME architecture for adjusting the parameters and the iteratively reweighted least squares (IRLS) algorithm is used to perform maximization in the inner loop [Jordan, M.I., Jacobs, R.A. (1994). Hierarchical mixture of experts and the EM algorithm, Neural Computation, 6(2), 181--214]. However, it is reported in literature that the IRLS algorithm is of instability and the ME architecture trained by the EM algorithm, where IRLS algorithm is used in the inner loop, often produces the poor performance in multiclass classification. In this paper, the reason of this instability is explored. We find out that due to an implicitly imposed incorrect assumption on parameter independence in multiclass classification, an incomplete Hessian matrix is used in that IRLS algorithm. Based on this finding, we apply the Newton--Raphson met...

In this paper we describe some algorithms for noisy optimization in the context of problems from the gas transmission industry. The algorithms are implicit filtering, DIRECT, and a new hybrid of these methods, which uses DIRECT to find an initial iterate for implicit filtering. We report on numerical results that illustrate the performance of the methods.

The goal of this article is to give an overview of numerical problems encountered when determining the electronic structure of materials and the rich variety of techniques used to solve these problems. The paper is intended for a diverse scienti£c computing audience. For this reason, we assume the reader does not have an extensive background in the related physics. Our overview focuses on the nature of the numerical problems to be solved, their origin, and on the methods used to solve the resulting linear algebra or nonlinear optimization problems. It is common knowledge that the behavior of matter at the nanoscale is, in principle, entirely determined by the Schrödinger equation. In practice, this equation in its original form is not tractable. Successful, but approximate, versions of this equation, which allow one to study nontrivial systems, took about £ve or six decades to develop. In particular, the last two decades saw a ¤urry of activity in developing effective software. One of the main practical variants of the Schrödinger equation is based on what is referred to as Density Functional Theory (DFT). The combination of DFT with pseudopotentials allows one to obtain in an ef£cient way the ground state con£guration for many materials. This article will emphasize pseudopotentialdensity

Let (x n ) and (x n ) be two vector sequences converging to a common limit. First, we shall define nonlinear hybrid procedures which consist of constructing a new vector sequence (y n ) with better convergence properties than (x n ) and (x n ). Then, this procedure is used for accelerating the convergence of a given sequence and applied to the construction of fixed point methods. New methods are derived. Finally, the connection between fixed point iterations and methods for the numerical integration of differential equations is also exploited. Numerical results are given.

bundle filter method for nonsmooth nonlinear

Finding optimal three-dimensional molecular configurations based on a limited amount of experimental and/or theoretical data requires efficient nonlinear optimization algorithms. Optimization methods must be able to find atomic configurations that are close to the absolute, or global, minimum error and also satisfy known physical constraints such as minimum separation distances between atoms (based on van der Waals interactions). The most difficult obstacles in these types of problems are that 1) using a limited amount of input data leads to many possible local optima and 2) introducing physical constraints, such as minimum separation distances, helps to limit the search space but often makes convergence to a global minimum more difficult. We introduce a constrained global optimization algorithm that is robust and efficient in yielding near-optimal three-dimensional configurations that are guaranteed to satisfy known separation constraints. The algorithm uses an atom-based approach that reduces the dimensionality and allows for tractable enforcement of constraints while maintaining good global convergence properties. We evaluate the new optimization algorithm using synthetic data from the yeast phenylalanine tRNA and several proteins, all with known crystal structure taken from the Protein Data Bank. We compare the results to commonly applied optimization methods, such as distance geometry, simulated annealing, continuation, and smoothing. We show that compared to other optimization approaches, our algorithm is able combine sparse input data with physical constraints in an efficient manner to yield structures with lower root mean squared deviation.

. In this paper we show how the implicit filtering algorithm can be parallelized and applied to problems in parameter identification and optimization in automotive valve train design. We extend our previous work by using a more refined model of the valve train and exploiting parallelism in a new way. We apply the parameter identification results to obtain optimal profiles for camshaft lobes. Key words. Noisy Optimization, Implicit Filtering, Mechanical Systems, Automotive Valve Trains AMS subject classifications. 65K05, 65K10, 65L05, 65Y05 1. Introduciton. In this paper we report on a parallel implementation of the implicit filtering [17], [19] algorithm and its application to problems in parameter identification and optimization in automotive valve train design. We extend our previous work [11], [10] on parameter identification by using a more refined model of the valve train and exploiting parallelism in a new way. We then apply the parameter identification results to obtain optim...

The topic of skeletal animation and its associated techniques have previously been applied in the area of animating computer-generated characters for motion pictures and computer games. This thesis investigates the use of similar techniques in the scope of exploring three-dimensional visualisations of relational networks (graphs). A system