English

The Gradient descent method from the perspective of fractional calculus

Optimization and Control 2020-12-22 v2

Abstract

Motivated by gradient methods in optimization theory, we give methods based on ψ\psi-fractional derivatives of order α\alpha in order to solve unconstrained optimization problems. The convergence of these methods is analyzed in detail. This paper also presents an Adams-Bashforth-Moulton (ABM) method for the estimation of solutions to equations involving ψ\psi-fractional derivatives. Numerical examples using the ABM method show that the fractional order α\alpha and weight ψ\psi are tunable parameters, which can be helpful for improving the performance of gradient descent methods.

Keywords

Cite

@article{arxiv.1911.08058,
  title  = {The Gradient descent method from the perspective of fractional calculus},
  author = {Pham Viet Hai and Joel A. Rosenfeld},
  journal= {arXiv preprint arXiv:1911.08058},
  year   = {2020}
}

Comments

27 pages