Feeding the fish - weight update strategies for the fish school search algorithm

Authors:
Andreas Janecek;Ying Tan
Affiliations:
Key Laboratory of Machine Perception (MOE), Peking University, Department of Machine Intelligence, School of Electronics Engineering and Computer Science, Peking University, Beijing, China;Key Laboratory of Machine Perception (MOE), Peking University, Department of Machine Intelligence, School of Electronics Engineering and Computer Science, Peking University, Beijing, China
Venue:
ICSI'11 Proceedings of the Second international conference on Advances in swarm intelligence - Volume Part II
Year:
2011

Citing 4
Cited 1

Genetic Algorithms in Search, Optimization and Machine Learning

Genetic Algorithms in Search, Optimization and Machine Learning
Differential Evolution – A Simple and Efficient Heuristic for Global Optimization over Continuous Spaces

Journal of Global Optimization
Using population based algorithms for initializing nonnegative matrix factorization

ICSI'11 Proceedings of the Second international conference on Advances in swarm intelligence - Volume Part II
Fireworks algorithm for optimization

ICSI'10 Proceedings of the First international conference on Advances in Swarm Intelligence - Volume Part I

Using population based algorithms for initializing nonnegative matrix factorization

ICSI'11 Proceedings of the Second international conference on Advances in swarm intelligence - Volume Part II

Quantified Score

Hi-index	0.00

Visualization

Abstract

Choosing optimal parameter settings and update strategies is a key issue for almost all population based optimization algorithms based on swarm intelligence. For state-of-the-art optimization algorithms the optimal parameter settings and update strategies for different problem sizes are well known. In this paper we investigate and compare different newly developed weight update strategies for the recently developed Fish School Search (FSS) algorithm. For this algorithm the optimal update strategies have not been investigated so far. We introduce a new dilation multiplier as well as different weight update steps where fish in poor regions loose weight more quickly than fish in regions with a lot of food. Moreover, we show how a simple non-linear decrease of the individual and volitive step parameters is able to significantly speed up the convergence of FSS.