Three talks: SIAM OP14 preview
1: A General Krylov Subspace Method and its Connection to the Conjugate-gradient Method
2: Generation of Relevant Elementary Flux Modes in Metabolic Networks,
3: Using Optimization Methods for Efficient Fatigue Criterion Evaluation
Time: Fri 2014-05-16 11.00 - 12.00
Location: Room 3721, Lindstedtsvägen 25, 7th floor, KTH
Participating: Tove Odland, Hildur Æsa Oddsdóttir, Henrik Svärd
Welcome to informal seminar in optimization and systems theory, Friday May 16, 11.00-12.00 in room 3721
Three talks will be given by PhD students at the division. These talks will be given next week at the SIAM Conference on Optimization 2014 in San Diego. (The length of the talks reflects the actual length at the SIAM Conference.)
11.00-11.25 Tove Odland: A General Krylov Subspace Method and its Connection to the Conjugate-gradient Method
Krylov subspace methods are used for solving systems of linear equations Ax=b. We present a general Krylov subspace method that can be used when the matrix A is indefinite. We show that the approximate solution in each step, xk, is given by the chosen scaling of the orthogonal vectors that successively make the Krylov subspaces available. Our framework gives an intuitive way to see the conjugate-gradient method and why it sometimes fails if A is not positive definite.
11.30-11.45 Hildur Æsa Oddsdóttir Generation of Relevant Elementary Flux Modes in Metabolic Networks
Elementary flux modes (EFMs) are a set of pathways in a metabolic reaction network that can be combined to define any feasible pathway in the network. It is in general prohibitive to enumerate all EFMs for complex networks. We present a method based on a column generation technique that finds EFMs in a dynamic fashion, while minimizing the norm of the difference between extracellular rate measurements and calculated flux through the EFMs in the network.
11.50-12.05 Henrik Svärd: Using Optimization Methods for Efficient Fatigue Criterion Evaluation
In many industrial fields, computer simulations of durability are increasing in importance. In such simulations, “critical plane” criteria are often used to determine the probability of fatigue crack initiation in a point. This type of criterion involves searching, in each material point, for the plane where the combination of stresses is the most damaging. This talk concerns new efficient ways to evaluate this type of criterion, which are based on the use of optimization methods, shortening the lead time and increasing the accuracy of the results.