الفهرس 
Inference Mechanism
An Enhanced Differential Evolution Optimization Algorithm: the Proposed ApproachOnly 14 pages are availabe for public view
 |  | from | 24 |  |  |
| | | from | 24 | | |
Abstract
The research point we have investigated is one of enhancing the differential evolution algorithm such that its performance is improved in terms of accuracy and convergence speed. The enhanced algorithm is to be applied to training ANFIS parameters. The differential evolution optimization algorithm (DE) is a population-based algorithm with three control parameters: population size (NP), mutation scaling factor (F), and crossover rate (CR). We proposed an enhanced DE algorithm and abbreviate it as ModDE. The resulting algorithm improves the performance of DE and simplifies the tuning for the control parameters. The control parameters of ModDE are reduced to only two control parameter (NP and F) instead of three in the classical DE (NP, F and CR). ModDE is tested on a total of 47 benchmark functions: 27 traditional functions and 20 special functions chosen from CEC2005 and CEC2013.The results are represented in terms of the mean and standard deviation of the error, success rate, and average number of function evaluations over successful runs. The convergence characteristics of the algorithm are studied via the variation of the best error value with the number of function evaluations, FEs. The results for the majority of the functions considered are acceptable and promising.