| 引用本文： |
-
程文旗,郭华,谢承旺,韦伟,潘嘉敏,龙广林.一种增强多样性的改进型NSGAⅡ算法[J].广西科学,2021,28(4):353-362. [点击复制]
- CHENG Wenqi,GUO Hua,XIE Chengwang,WEI Wei,PAN Jiamin,LONG Guanglin.An Improved NSGAⅡ Algorithm for Enhancing Diversity[J].Guangxi Sciences,2021,28(4):353-362. [点击复制]
|
|
| 本文已被：浏览 376次 下载 453次 |
码上扫一扫！ |
| 一种增强多样性的改进型NSGAⅡ算法 |
|
程文旗1, 郭华1, 谢承旺1,2, 韦伟1, 潘嘉敏1, 龙广林1
|
|
|
| (1.南宁师范大学, 计算机与信息工程学院, 广西南宁 530000;2.华南师范大学, 数据科学与工程学院, 广东汕尾 516600) |
|
| 摘要: |
| 传统NSGAⅡ算法通过计算个体的拥挤距离保持群体分布性。为改善算法中存在的不足，使得新算法在测试问题集上表现更好，本研究对算法的多样性进行改进。受PBI效用函数的启发，抽取其中的d2距离作为选择机制并与传统NSGAⅡ算法结合，提出一种计算d2距离的改进型NSGAⅡ算法（d2_NSGAⅡ），用于改善传统算法的收敛性与多样性。通过实验对比发现，相比NSGAⅡ以及其他一些算法，新算法在一些测试函数的高维多目标优化问题上有明显的优势。因此，d2_NSGAⅡ是一种较好的解决高维多目标优化问题的新算法。 |
| 关键词: 多目标优化 非支配排序 进化算法 拥挤距离 NSGAⅡ |
| DOI：10.13656/j.cnki.gxkx.20211109.006 |
| 投稿时间：2021-05-28 |
| 基金项目:国家自然科学基金（61763010），广西自然科学基金（2021GXNSFAA075011）和广西研究生教育创新计划项目（YCSW2020194）资助。 |
|
| An Improved NSGAⅡ Algorithm for Enhancing Diversity |
|
CHENG Wenqi1, GUO Hua1, XIE Chengwang1,2, WEI Wei1, PAN Jiamin1, LONG Guanglin1
|
| (1.School of Computer and Information Engineering, Nanning Normal University, Nanning, Guangxi, 530000, China;2.School of Data Science & Engineering, South China Normal University, Shanwei, Guangdong, 516600, China) |
| Abstract: |
| The traditional NSGAⅡ algorithm maintains the population distribution by calculating the crowding distance of individuals. In order to improve the shortcomings of the algorithm and make the new algorithm perform better on the test problem set, the diversity of the algorithm is improved in this study. Inspired by the PBI utility function, the d2 distance is extracted as the selection mechanism and combined with the traditional NSGAⅡ algorithm. An improved NSGAⅡ algorithm for calculating d2 distance (d2_ NSGAⅡ) is proposed to improve the convergence and diversity of traditional algorithms. Through experimental comparison, it is found that compared with NSGAⅡ and other algorithms, the new algorithm has obvious advantages in high-dimensional multi-objective optimization problems for some test functions.Therefore, d2_NSGAⅡ is a better new algorithm to solve high-dimensional multi-objective optimization problems. |
| Key words: multi-objective optimization non-dominated sort evolutionary algorithm crowding distance NSGAⅡ |
