A robust and efficient combined trust region–line search approach for constrained nonlinear least squares problems

Speaker: Prof. Nezam Mahdavi-Amiri
Faculty of Mathematical Sciences
Sharif University of Technology

Title: A robust and efficient combined trust region–line search approach for constrained nonlinear least squares problems

Date and time: Friday 7 July 2017, 3:00–4:00pm
Location: Building 8 Level 9 Room 66 (AGR) RMIT City campus

Abstract: We describe a combined trust region–line search projected structured algorithm for solving constrained nonlinear least-squares problems. The approach is based on an adaptive projected structured scheme due to Mahdavi-Amiri and Bartels based on an exact penalty method and has been shown to have a local two-step superlinear rate of convergence. For robustness, a new penalty parameter updating strategy and a specific line search technique within the trust region are employed. Technical details of our implementation are discussed and the program is tested on known test problems (small and large residuals) and some randomly generated ones. A comparison of our obtained results with the ones obtained by a number of well-known general nonlinear programming methods, while showing competitiveness of the algorithm, confirms the practical significance of the adaptive penalty updating scheme, combined trust region–line search strategy, and the special structured consideration for the approximate projected least squares Hessian.

Bio: Nezam Mahdavi-Amiri is a Distinguished Professor at Sharif University of Technology, Iran, and President of the Iranian Operations Research Society. He received Bachelor degree in Computer Science from Louisiana State University, USA, Master degrees in Computer Science from Pennsylvania State University, USA and in Mathematical Sciences from Johns Hopkins University, USA, and PhD degree from Johns Hopkins University, USA. To date, he has published more than 230 papers, supervised 20 PhD students and delivered 28 invited talks in international conferences and workshops in optimisation and its applications. Professor Mahdavi-Amiri’s research interests include numerical optimisation, fuzzy optimisation, numerical linear algebra, mathematical software, mathematical modelling and data fitting.

Please note that Prof. Nezam Mahdavi-Amiri will be giving a talk at CIAO (Federation University Australia) on 6 July. Visimeet connection will be available. Please contact Alex Kruger form more information.

Download Poster

Skip to toolbar