Summary
International Symposium on Nonlinear Theory and Its Applications
2022
Session Number:A2L-D
Session:
Number:A2L-D-04
Herding with Self-Organizing Multiple Starting Point Optimization
Hiroshi Yamashita , Hideyuki Suzuki , Kazuyuki Aihara,
pp.33-36
Publication Date:12/12/2022
Online ISSN:2188-5079
DOI:10.34385/proc.71.A2L-D-04
PDF download (475.2KB)
Summary:
The herding algorithm is a prominent sampling algorithm using the complex behavior of high-dimensional nonlinear dynamics. Its procedure is composed of the step of nonconvex optimization and that of updating the state depending on its output. This completely deterministic algorithm is interesting as it connects optimization algorithm, nonlinear dynamics and sampling. In this paper, we propose a multiple starting point heuristic for the optimization step of herding and discuss the behavior of the algorithm that is also worth pursuing in itself as a nonlinear dynamical system. In particular, we observe that the candidate states in the algorithm keep distance from each other, although it is not designed explicitly.