Title :
Using Accelerator Feedback to Improve Performance of Integral-Controller Particle Swarm Optimization
Author :
Zhihua Cui ; Zeng, Jianchao ; Sun, Guoji
Author_Institution :
State Key Lab. for Manuf. Syst. Eng., Xi´´an Jiaotong Univ.
Abstract :
Integral-controller particle swarm optimization (ICPSO), influenced by inertia weight w and coefficient phi is a new swarm technology by adding accelerator information. Based on stability analysis, the convergence conditions imply the negative selection principles of inertia weight w, and the relationship between w and phi. To improve the computational efficiency, an adaptive strategy for tuning the parameters of ICPSO is described using a new statistical variable reflecting computational efficiency index-average accelerator information. The optimization computing of some examples is made to show that the ICPSO has better global search capacity and rapid convergence speed
Keywords :
convergence; particle swarm optimisation; search problems; stability; statistics; accelerator feedback; adaptive strategy; convergence conditions; global searching; integral-controller particle swarm optimization; negative selection principles; parameter tuning; stability analysis; statistics; Computational efficiency; Control systems; Convergence; Integral equations; Laboratories; Particle accelerators; Particle swarm optimization; Stability analysis; State feedback; Sun; Average Accelerator Information; Integral-controller; Particle swarm optimization;
Conference_Titel :
Cognitive Informatics, 2006. ICCI 2006. 5th IEEE International Conference on
Conference_Location :
Beijing
Print_ISBN :
1-4244-0475-4
DOI :
10.1109/COGINF.2006.365567
