语义关系算子在群智能算法中的研究

STUDY ON SEMANTIC RELATION OPERATORS IN SWARM INTELLIGENCE ALGORITHM

在传统群智能算法框架的基础上,提出基于语义关系算子的群智能算法。与传统的群智能算法不同,该算法采用了一个语义关系算子进行关键参数的更新操作,并基于群体的历史状态对语义关系可进行更新。在语义关系库更新过程中,通过分析群体之前的历史状态信息,进行基于本体的语义关系挖掘,从而找到全局语义关系。以粒子群算法解组合优化问题为例,提出了基于语义关系算子的蚁群算法和粒子群算法。实验表明,基于该算子的群智能算法寻优能力有了一定程度的改进。

Built on the traditional swarm intelligence algorithm framework, a semantic relation operator based swarm intelligence algorithm is proposed. In contrast with the tradition swarm intelligence algorithm, the proposed one uses semantic relation operators to update crucial parameters. In addition, based on swarm historical status, the semantic relation can be updated too. During semantic relation library update, through analyzing swarm＇s previous historical status information, it executes ontology-based semantic relation mining in order to find out global semantic relations. Taking particle swarm algorithm for solving combined optimization problems as example, an ant colony algorithm and a particle swarm algorithm, both of which are based on semantic relation operators, are proposed. Experiment shows there are improvements in optimum searching capability with the operator-based swarm intelligence algorithm.