Locust Swarms are a newly developed multi-optima particle swarm. They were explicitly developed for non-globally convex search spaces, and their non-convergent search behaviours can also be useful for problems with fractal and fractured landscapes. On the 1000-dimensional "FastFractal" problem used in the 2008 CEC competition on Large Scale Global Optimization, Locust Swarms can perform better than all of the methods in the competition. Locust Swarms also perform very well on a real-world optimization problem that has a fractured landscape. The extent and the effects of a fractured landscape are observed with a practical new measurement that is affected by the degree of fracture and the lack of regularity and symmetry in a fitness landscape. © Springer-Verlag Berlin Heidelberg 2009.
CITATION STYLE
Chen, S., & Lupien, V. (2009). Optimization in fractal and fractured landscapes using locust swarms. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5865 LNAI, pp. 232–241). https://doi.org/10.1007/978-3-642-10427-5_23
Mendeley helps you to discover research relevant for your work.