|
|
|
| A HYBRID GENETIC ALGORITHM FOR ESTIMATION OF PARAMETER OF NONLINEAR REGRESSION MODELINGS |
| Chen Jinshan; Wei Gang |
| College of Electron. Info., South China Univ. of Technology Guangzhou 510641 China |
|
|
|
|
Abstract A new hybrid genetic algorithm (HGA) that incorporates the flexible polyhedron method into the conventional genetic algorithm is proposed. The HGA speeds up convergence before the genetic algorithm reachs the global optima, and achieves better performance by performing global search and local search alternatively. The experiments for parameter estimation of nonlinear regression modelings show that the HGA has better generalization and effectiveness.
|
|
Received: 23 December 1999
|
|
|
|
|
|
|
|
|
2001, Vol. 23 
