Wed.3 16:00–17:15 | H 3007 | NON
.

Recent Trends in Large Scale Nonlinear Optimization (2/2)

Chair: Panos Parpas Organizers: Michal Kočvara, Panos Parpas
16:00

Filippo Pecci

joint work with Panos Parpas, Ivan Stoianov

Non-linear inverse problems via sequential convex optimization

We focus on inverse problems for physical systems subject to a set of non-linear equations and formulate a generic sequential convex optimization solution method. The proposed algorithm is implemented for fault estimation in water supply networks. Previous literature has solved this combinatorial problem using evolutionary approaches. We show that the considered problem can be approximated by a non-linear inverse problem with L1 regularization and solved using sequential convex optimization. The developed method is tested using a large operational water network as case study.

16:25

Michal Kočvara

joint work with Panos Parpas, Chin Pang Ho

On Newton-type multilevel optimization methods

Building upon multigrid methods, a framework of multilevel optimization methods was developed to solve structured optimization problems, including problems in optimal control, image processing, etc. In this talk, we give a broader view of the multilevel framework and establish some connections between multilevel algorithms and the other approaches. By studying a case study of the so-called Newton-type multilevel model, we take the first step to show how the structure of optimization problems could improve the convergence of multilevel algorithms.

16:50

Man Shun Ang

joint work with Nicolas Gillis

Accelerating Nonnegative-X by extrapolation, where X ∈ {Least Square, Matrix Factorization, Tensor Factorization}

Non-negativity constraint is ubiquitous. A family of non-negativity constrained problems (NCP) are the non-negative least squares (NNLS), the non-negative matrix factorization (NMF) and the non-negative tensor factorization (NTF). The goal of these NCP problems in general is to fit an given object (vector, matrix, or tensor) by solving an optimization problem with non-negativity constraints. Various algorithms are proposed to solve the NCP. In this talk, we introduce an general extrapolation-based acceleration scheme that that can significantly accelerate the algorithm for NNLS, NMF and NTF.