In this paper we propose a novel hybrid version of Differential Evolution (DE). Firstly we modify the traditional structure of population in DE and propose a new strategy for population setting, in which the population is sorted in line with the fitness values of individuals. Another method is saltatory sampling with a nonrandom order, which is utilized to select candidates for the mutation operation. Furthermore, a strategy of survival of the fittest was used for individual selection operation. Combined the two strategies with DE, differential evolution based on individual-sorting and individual-sampling (ISSDE) is proposed, via testing on a series benchmark functions and compared with three variants of DE, the simulation results show that the proposed ISSDE has a better performance both in convergence speed and robustness. 1553-9105/
|Number of pages||9|
|Journal||Journal of Computational Information Systems|
|Publication status||Published - Jan 2012|
Bibliographical noteThis research was supported by the Natural Science Foundation of Zhejiang Province under Grant No. Y1100076, K. C. Wong Magna Fund in Ningbo University, and the Scientific Research Foundation of Graduate School of Ningbo University.
- Differential evolution