2010 | OriginalPaper | Buchkapitel
Improving Local Convergence in Particle Swarms by Fitness Approximation Using Regression
verfasst von : Stefan Bird, Xiaodong Li
Erschienen in: Computational Intelligence in Expensive Optimization Problems
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
In this chapter we present a technique that helps Particle Swarm Optimisers (PSOs) locate an optimum more quickly, through fitness approximation using regression. A least-squares regression is used to estimate the shape of the local fitness landscape. From this shape, the expected location of the peak is calculated and the information given to the PSO. By guiding the PSO to the optimum, the local convergence speed can be vastly improved. We demonstrate the effectiveness of using regression on several static multimodal test functions as well as dynamicmultimodal test scenarios (Moving Peaks). This chapter also extends theMoving Peaks test suite by enhancing the standard conic peak function to allow the creation of asymmetrical and additional multiple local peaks. The combination of this technique and a speciation-based PSO compares favourably to another multi-swarm PSO algorithm that has proven to be working well on the Moving peaks test functions.