Cellular Differential Evolution for Optimization
This paper is to illustrate the Cellular Differential Evolution with the cellular structure originated from Cellular automata. Cellular neighbor local search has been designed; base vector or global best in mutation operator is substituted by neighborhood-best, which overcomes the weakness of single selection relating to global best, and balances the contradiction of local and global search, and improves the diversity of population. In addition, cellular structure ensures information exchange, inheritance and diffusion. Finally, a specific algorithm has been implemented: compared with similar variants of DE, the simulation results on 9 benchmark functions demonstrate that cellular differential evolutions are provided with obvious advantages in the solution-quality, stability and speed.
G. W. Hu and X. Y. Du, "Cellular Differential Evolution for Optimization", Key Engineering Materials, Vols. 474-476, pp. 1770-1775, 2011