کد مقاله | کد نشریه | سال انتشار | مقاله انگلیسی | نسخه تمام متن |
---|---|---|---|---|
495296 | 862822 | 2015 | 17 صفحه PDF | دانلود رایگان |
• A novel optimisation algorithm, named enhanced leader PSO (ELPSO), is introduced.
• ELPSO mitigates premature convergence problem of conventional PSO.
• ELPSO is mainly based on a five-staged successive mutation strategy.
• At each iteration, the successive mutation strategy is applied to swarm leader.
• The results confirm the outperformance of ELPSO over other compared algorithms.
Particle swarm optimisation (PSO) is a well-established optimisation algorithm inspired from flocking behaviour of birds. The big problem in PSO is that it suffers from premature convergence, that is, in complex optimisation problems, it may easily get trapped in local optima. In this paper, a new PSO variant, named as enhanced leader PSO (ELPSO), is proposed for mitigating premature convergence problem. ELPSO is mainly based on a five-staged successive mutation strategy which is applied to swarm leader at each iteration. The experimental results confirm that in all terms of accuracy, scalability and convergence rate, ELPSO performs well.
Figure optionsDownload as PowerPoint slide
Journal: Applied Soft Computing - Volume 26, January 2015, Pages 401–417