Qualitative and quantitative assessment of step size adaptation rules

Publikation: Bidrag til bog/antologi/rapportKonferencebidrag i proceedingsForskningfagfællebedømt

We present a comparison of step size adaptation methods for evolution strategies, covering recent developments in the field. Following recent work by Hansen et al. we formulate a concise list of performance criteria: a) fast convergence of the mean, b) near-optimal fixed point of the normalized step size dynamics, and c) invariance to adding constant dimensions of the objective function. Our results show that algorithms violating these principles tend to underestimate the step size or are unreliable when the function does not fit to the algorithm's tuned hyperparameters. In contrast, we find that cumulative step size adaptation (CSA) and twopoint adaptation (TPA) provide reliable estimates of the optimal step size. We further find that removing the evolution path of CSA still leads to a reliable algorithm without the computational requirements of CSA.

TitelProceedings of the 14th ACM/SIGEVO Conference on Foundations of Genetic Algorithms
Antal sider10
ForlagAssociation for Computing Machinery
ISBN (Elektronisk)978-1-4503-4651-1
StatusUdgivet - 2017
Begivenhed14th ACM/SIGEVO Conference on Foundations of Genetic Algorithms - København, Danmark
Varighed: 12 jan. 201715 jan. 2017
Konferencens nummer: 14


Konference14th ACM/SIGEVO Conference on Foundations of Genetic Algorithms

ID: 179557726