An Improved APSO-SQP with Adaptive Transition Strategy for Dynamic Optimization

fang dong,jianming zhang,lei xie,huining zhao,xiongxiong he
DOI: https://doi.org/10.3182/20140824-6-za-1003.00580
2014-01-01
IFAC Proceedings Volumes
Abstract:Abstract This paper employs a logistical regression classifier for combining particle swarm optimizer and local optimizer to solve dynamic optimization problems. It achieves the ability of automatic transition from global optimizer to local optimizer according to the state of solutions. The particle swarm optimizer is used to globally search for a good approximation of the solution and then SQP is applied as the local optimizer for quickly converging to the optima. Experimental results on two complex chemical processes demonstrate that the proposed method can get comparable and even better results than the existed methods in precision and especially in function evaluations.
What problem does this paper attempt to address?