基于免疫关联规则的烟气含氧量优化研究

Optimization of Flue Gas Oxygen Content Based on Immune Association Rules

给出了一种基于矢量距的自适应免疫遗传算法的多维关联规则挖掘算法。该算法在简单遗传算法的基础上引入免疫算子和自适应交叉、变异算子,并充分利用了人工免疫的记忆特性,把挖掘的关联规则存入记忆库,加快了关联规则的挖掘速度。并将该算法应用于电站锅炉燃烧系统最佳烟气含氧量的研究,阐述了研究结果。

A mining algorithm for multi-dimensional association rule is proposed based on adaptive immune genetic algorithm with vector distance. The algorithm integrates immune operator, adaptive crossover operator and mutation operator with simple genetic algorithm. Besides, it makes full use of artificial immune memory characteristics, and stores the mined association rules in memory, and accelerates the mining rate. The algorithm has been applied in the research of the optimal flue gas oxygen content in the power plants boiler combustion system. As the application resuits indicate, the proposed method can obtain potential rules and trends from the enormous real-time historical data, and conduct global optimization search quickly and effectively, especially feasible for large-scale, massive database mining.