Abstract
Section 1 of the full paper briefs the principles of PF, including the four steps of PF algorithm. Section 2 explains how we combine dynamic EA with PF; we call our algorithm PFDGA (Particle Filter Based Dynamic Genetic Algorithm) and present a 16-step procedure for obtaining PFDGA; its core consists of: (1) we use PF to predict the changing tendency of the optimal solution of a dynamic optimization problem in the decision space; (2) we propose a self-adaptive population diversity control method to balance the effects of EA and PF. Section 3 presents experimental results, which are given in Figs. 2 through 7, and their analysis; in section 3, using the moving peak benchmark (MPB) problem, we compare the performance of PFDGA with that of random immigrant genetic algorithm (RIGA). The experimental results and their analysis show preliminarily that our PFDGA is efficient and can solve dynamic optimization problems more effectively.
Original language | English |
---|---|
Pages (from-to) | 393-397 |
Number of pages | 5 |
Journal | Xibei Gongye Daxue Xuebao/Journal of Northwestern Polytechnical University |
Volume | 28 |
Issue number | 3 |
State | Published - Jun 2010 |
Keywords
- Algorithms
- Dynamic evolutionary algorithm (EA)
- Dynamic optimization problem
- Optimization
- Particle filter (PF)