Home // ADVCOMP 2015, The Ninth International Conference on Advanced Engineering Computing and Applications in Sciences // View article
Authors:
Tanja Clees
Nils Hornung
Igor Nikitin
Lialia Nikitina
Daniela Steffes-lai
Keywords: complex computing in application domains; medical computation and graphics; advanced computing in simulation systems; advanced computing for statistics and optimization
Abstract:
We consider bi-objective optimization problem from noninvasive tumor therapy planning. The therapy uses magnetic resonance tomography for the location of the target region and focused ultrasound for the destruction of tumor cells. Experimentally validated physical models are used to construct numerical simulation including nonlinear wave propagation, absorption in soft tissue, heat transfer and a hierarchical structure of the biological materials. The resulting cumulative thermal dose inside the target region should be maximized, providing a maximal level of tumor destruction, while the thermal dose outside the target region should be minimized, to decrease the influence to healthy organs. Metamodeling with radial basis functions is used for continuous representation of optimization objectives. The problem possesses nonconvex Pareto front. Detection of nonconvex Pareto fronts is especially difficult, this is a point where many simple algorithms fail. In this paper we consider different approaches to this problem: sequential linear programming (SLP), sequential quadratic programming (SQP) and generic 1- or 2-phase nonlinear programming (NLP). We show the ability of the algorithms to process such case and compare the efficiency of different approaches.
Pages: 71 to 76
Copyright: Copyright (c) IARIA, 2015
Publication date: July 19, 2015
Published in: conference
ISSN: 2308-4499
ISBN: 978-1-61208-419-0
Location: Nice,France
Dates: from July 19, 2015 to July 24, 2015