Abstract: Aiming at the low efficiency of existing interval multi-objective optimization methods,a fast interval multi-objective optimization method based on fuzzy rule base was proposed.In this method,Autonomous Learning Multi-Model(ALMMo)system was used to learn fuzzy rules,and the fuzzy rule base was formed to represent interval objective function and interval constraint possibility degree,which could replace inner layer optimization and speed up the optimization process.In addition,in the candidate solution set generated by each generation of fuzzy rule base,the appropriate individuals were selected for real evaluation and the fuzzy rule base was updated adaptively,which improved the optimization accuracy and computational efficiency and reduced the system complexity.The proposed method was applied to four numerical cases and an engineering example.Compared with the existing representative methods,the proposed method had higher computational efficiency without losing accuracy.

Key words: uncertain optimization, interval optimization, interval multi-objective optimization, fuzzy rule base, autonomous learning multi-model

摘要： 针对现有区间多目标优化方法低效的问题,提出一种基于模糊规则库的快速区间多目标优化方法。该方法采用自主学习多模型系统学习模糊规则,并组成模糊规则库以表示区间目标函数和区间约束可能度,替代内层优化,加快优化进程;在每一代模糊规则库生成的备选解集中选择适量个体进行真实评估并自主更新模糊规则库,提高了优化精度和计算效率,降低了系统复杂性。将所提方法运用在4个数值案例和1个工程实例,与现有的代表性方法相比,所提方法在不失精度的前提下具有更高的计算效率。

关键词: 不确定优化, 区间优化, 区间多目标优化, 模糊规则库, 自主学习多模型