TWO NEW WEAK CONSTRAINT QUALIFICATIONS AND APPLICATIONS

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 the assumption that many subsets of the gradients of the active constraints preserve positive linear dependence locally. A major open question was to identify the exact set of gradients whose properties had to be preserved locally and that would still work as a CQ. This is done in the first new CQ, which we call the constant rank of the subspace component (CRSC) CQ. This new CQ also preserves many of the good properties of RCPLD, such as local stability and the validity of an error bound. We also introduce an even weaker CQ, called the constant positive generator (CPG), which can replace RCPLD in the analysis of the global convergence of algorithms. We close this work by extending convergence results of algorithms belonging to all the main classes of nonlinear optimization methods: sequential quadratic programming, augmented Lagrangians, interior point algorithms, and inexact restoration.

Black-scholes type equations

In this work we are concerned with the analysis and numerical solution of Black-Scholes type equations arising in the modeling of incomplete financial markets and an inverse problem of determining the local volatility function in a generalized Black-Scholes model from observed option prices. In the first chapter a fully nonlinear Black-Scholes equation which models transaction costs arising in option pricing is discretized by a new high order compact scheme. The compact scheme is proved to be unconditionally stable and non-oscillatory and is very efficient compared to classical schemes. Moreover, it is shown that the finite difference solution converges locally uniformly to the unique viscosity solution of the continuous equation. In the next chapter we turn to the calibration problem of computing local volatility functions from market data in a generalized Black-Scholes setting. We follow an optimal control approach in a Lagrangian framework. We show the existence of a global solution and study first- and second-order optimality conditions. Furthermore, we propose an algorithm that is based on a globalized sequential quadratic programming method and a primal-dual active set strategy, and present numerical results. In the last chapter we consider a quasilinear parabolic equation with quadratic gradient terms, which arises in the modeling of an optimal portfolio in incomplete markets. The existence of weak solutions is shown by considering a sequence of approximate solutions. The main difficulty of the proof is to infer the strong convergence of the sequence. Furthermore, we prove the uniqueness of weak solutions under a smallness condition on the derivatives of the covariance matrices with respect to the solution, but without additional regularity assumptions on the solution. The results are illustrated by a numerical example.

Uniform Convergence of the Newton Method for Aubin Continuous Maps

* This work was supported by National Science Foundation grant DMS 9404431.

Thrust allocation with linear constrained quadratic cost function

In this paper, a method of thrust allocation based on a linearly constrained quadratic cost function capable of handling rotating azimuths is presented. The problem formulation accounts for magnitude and rate constraints on both thruster forces and azimuth angles. The advantage of this formulation is that the solution can be found with a finite number of iterations for each time step. Experiments with a model ship are used to validate the thrust allocation system.

Optimum Load Shedding Through Programming Techniques

This paper obtains a new accurate model for sensitivity in power systems and uses it in conjunction with linear programming for the solution of load-shedding problems with a minimum loss of loads. For cases where the error in the sensitivity model increases, other linear programming and quadratic programming models have been developed, assuming currents at load buses as variables and not load powers. A weighted error criterion has been used to take priority schedule into account; it can be either a linear or a quadratic function of the errors, and depending upon the function appropriate programming techniques are to be employed.

Fast Eigen solution for homogeneous quadratic minimization with at most three constraints

We propose an eigenvalue based technique to solve the Homogeneous Quadratic Constrained Quadratic Programming problem (HQCQP) with at most three constraints which arise in many signal processing problems. Semi-Definite Relaxation (SDR) is the only known approach and is computationally intensive. We study the performance of the proposed fast eigen approach through simulations in the context of MIMO relays and show that the solution converges to the solution obtained using the SDR approach with significant reduction in complexity.

Lattice constellations and codes from quadratic number fields

We propose new classes of linear codes over integer rings of quadratic extensions of Q, the field of rational numbers. The codes are considered with respect to a Mannheim metric, which is a Manhattan metric modulo a two-dimensional (2-D) grid. In particular, codes over Gaussian integers and Eisenstein-Jacobi integers are extensively studied. Decoding algorithms are proposed for these codes when up to two coordinates of a transmitted code vector are affected by errors of arbitrary Mannheim weight. Moreover, we show that the proposed codes are maximum-distance separable (MDS), with respect to the Hamming distance. The practical interest in such Mannheim-metric codes is their use in coded modulation schemes based on quadrature amplitude modulation (QAM)-type constellations, for which neither the Hamming nor the Lee metric is appropriate.

Multiclass Core Vector Machine

Even though several techniques have been proposed in the literature for achieving multiclass classification using Support Vector Machine(SVM), the scalability aspect of these approaches to handle large data sets still needs much of exploration. Core Vector Machine(CVM) is a technique for scaling up a two class SVM to handle large data sets. In this paper we propose a Multiclass Core Vector Machine(MCVM). Here we formulate the multiclass SVM problem as a Quadratic Programming(QP) problem defining an SVM with vector valued output. This QP problem is then solved using the CVM technique to achieve scalability to handle large data sets. Experiments done with several large synthetic and real world data sets show that the proposed MCVM technique gives good generalization performance as that of SVM at a much lesser computational expense. Further, it is observed that MCVM scales well with the size of the data set.

Rough Core Vector Clustering

Support Vector Clustering has gained reasonable attention from the researchers in exploratory data analysis due to firm theoretical foundation in statistical learning theory. Hard Partitioning of the data set achieved by support vector clustering may not be acceptable in real world scenarios. Rough Support Vector Clustering is an extension of Support Vector Clustering to attain a soft partitioning of the data set. But the Quadratic Programming Problem involved in Rough Support Vector Clustering makes it computationally expensive to handle large datasets. In this paper, we propose Rough Core Vector Clustering algorithm which is a computationally efficient realization of Rough Support Vector Clustering. Here Rough Support Vector Clustering problem is formulated using an approximate Minimum Enclosing Ball problem and is solved using an approximate Minimum Enclosing Ball finding algorithm. Experiments done with several Large Multi class datasets such as Forest cover type, and other Multi class datasets taken from LIBSVM page shows that the proposed strategy is efficient, finds meaningful soft cluster abstractions which provide a superior generalization performance than the SVM classifier.

Infinite push: a new support vector ranking algorithm that directly optimizes accuracy at the absolute top of the list

Ranking problems have become increasingly important in machine learning and data mining in recent years, with applications ranging from information retrieval and recommender systems to computational biology and drug discovery. In this paper, we describe a new ranking algorithm that directly maximizes the number of relevant objects retrieved at the absolute top of the list. The algorithm is a support vector style algorithm, but due to the different objective, it no longer leads to a quadratic programming problem. Instead, the dual optimization problem involves l1, ∞ constraints; we solve this dual problem using the recent l1, ∞ projection method of Quattoni et al (2009). Our algorithm can be viewed as an l∞-norm extreme of the lp-norm based algorithm of Rudin (2009) (albeit in a support vector setting rather than a boosting setting); thus we refer to the algorithm as the ‘Infinite Push’. Experiments on real-world data sets confirm the algorithm’s focus on accuracy at the absolute top of the list.

Vehicle trajectory linearisation to enable efficient optimisation of the constant speed racing line

A driver model is presented capable of optimising the trajectory of a simple dynamic nonlinear vehicle, at constant forward speed, so that progression along a predefined track is maximised as a function of time. In doing so, the model is able to continually operate a vehicle at its lateral-handling limit, maximising vehicle performance. The technique used forms a part of the solution to the motor racing objective of minimising lap time. A new approach of formulating the minimum lap time problem is motivated by the need for a more computationally efficient and robust tool-set for understanding on-the-limit driving behaviour. This has been achieved through set point-dependent linearisation of the vehicle model and coupling the vehicle-track system using an intrinsic coordinate description. Through this, the geometric vehicle trajectory had been linearised relative to the track reference, leading to new path optimisation algorithm which can be formed as a computationally efficient convex quadratic programming problem. © 2012 Copyright Taylor and Francis Group, LLC.

Model based control for closed loop testing of 1-D engine simulation models

1-D engine simulation models are widely used for the analysis and verification of air-path design concepts and prediction of the resulting engine transient response. The latter often requires closed loop control over the model to ensure operation within physical limits and tracking of reference signals. For this purpose, a particular implementation of Model Predictive Control (MPC) based on a corresponding Mean Value Engine Model (MVEM) is reported here. The MVEM is linearised on-line at each operating point to allow for the formulation of quadratic programming (QP) problems, which are solved as the part of the proposed MPC algorithm. The MPC output is used to control a 1-D engine model. The closed loop performance of such a system is benchmarked against the solution of a related optimal control problem (OCP). As an example this study is focused on the transient response of a light-duty car Diesel engine. For the cases examined the proposed controller implementation gives a more systematic procedure than other ad-hoc approaches that require considerable tuning effort. © 2012 IFAC.

Output feedback H2 model matching for decentralized systems with delays

This paper gives a new solution to the output feedback H2 model matching problem for a large class of delayed information sharing patterns. Existing methods for similar problems typically reduce the decentralized problem to a centralized problem of higher state dimension. In contrast, this paper demonstrates that the decentralized model matching solution can be constructed from the original centralized solution via quadratic programming. © 2013 AACC American Automatic Control Council.

Plant-wide predictive control for a thermal power plant based on a physical plant model

A constrained non-linear, physical model-based, predictive control (NPMPC) strategy is developed for improved plant-wide control of a thermal power plant. The strategy makes use of successive linearisation and recursive state estimation using extended Kalman filtering to obtain a linear state-space model. The linear model and a quadratic programming routine are used to design a constrained long-range predictive controller One special feature is the careful selection of a specific set of plant model parameters for online estimation, to account for time-varying system characteristics resulting from major system disturbances and ageing. These parameters act as nonstationary stochastic states and help to provide sufficient degrees-of-freedom to obtain unbiased estimates of controlled outputs. A 14th order non-linear plant model, simulating the dominant characteristics of a 200 MW oil-fired pou er plant has been used to test the NPMPC algorithm. The control strategy gives impressive simulation results, during large system disturbances and extremely high rate of load changes, right across the operating range. These results compare favourably to those obtained with the state-space GPC method designed under similar conditions.

Improved conjugate gradient implementation for least squares support vector machines

As a promising method for pattern recognition and function estimation, least squares support vector machines (LS-SVM) express the training in terms of solving a linear system instead of a quadratic programming problem as for conventional support vector machines (SVM). In this paper, by using the information provided by the equality constraint, we transform the minimization problem with a single equality constraint in LS-SVM into an unconstrained minimization problem, then propose reduced formulations for LS-SVM. By introducing this transformation, the times of using conjugate gradient (CG) method, which is a greatly time-consuming step in obtaining the numerical solution, are reduced to one instead of two as proposed by Suykens et al. (1999). The comparison on computational speed of our method with the CG method proposed by Suykens et al. and the first order and second order SMO methods on several benchmark data sets shows a reduction of training time by up to 44%. (C) 2011 Elsevier B.V. All rights reserved.