Abstract
This paper demonstrates the optimality of an interpolation set employed in derivative-free trust-region methods. This set is optimal in the sense that it minimizes the constant of well-poisedness in a ball centred at the starting point. It is chosen as the default initial interpolation set by many derivative-free trust-region methods based on underdetermined quadratic interpolation, including NEWUOA, BOBYQA, LINCOA, and COBYQA. Our analysis provides a theoretical justification for this choice.
Acknowledgements
This paper corresponds to Section 2.5 of the PhD thesis of Tom M. Ragonneau [Citation24], co-supervised by Zaikun Zhang and Professor Xiaojun Chen from The Hong Kong Polytechnic University. Both authors are very grateful to Professor Chen for her support, encouragement, and guidance during the thesis. Zaikun Zhang would like to thank the late Professor Oleg Burdakov for his friendship.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 Powell's BOBYQA code contains a comment that ‘choices that exceed 2n + 1 are not recommended’.
Additional information
Funding
Notes on contributors
Tom M. Ragonneau
Tom M. Ragonneau obtained his PhD degree from The Hong Kong Polytechnic University in 2023. He is now a postdoctoral fellow of the CAS AMSS-PolyU Joint Laboratory of Applied Mathematics at The Hong Kong Polytechnic University.
Zaikun Zhang
Zaikun Zhang obtained his PhD degree from the Chinese Academy of Sciences in 2012. He worked at the University of Coimbra (Portugal) in 2012–2014 and then at Toulouse INP-ENSEEIHT (France) in 2014–2016 as a postdoctoral fellow. In 2016, he joined The Hong Kong Polytechnic University, and he is now an Assistant Professor at the Department of Applied Mathematics.