Performance boundary identification method for autonomous decision-making systems based on neighbor boundary degree
-
摘要:
性能边界是度量自主决策系统鲁棒性的重要表征，可以反映自主决策系统对抗扰动的能力。针对性能边界数据多空间分布、增量生成等特点，提出一种基于邻居边界度的自主决策系统性能边界识别方法。面对性能边界搜索空间复杂、全空间尺度不统一的难点，设计邻居边界度指标解决绝对尺度度量问题，并提出基于邻居边界度的性能边界识别流程；考虑利用增量数据结合原有识别结果进一步精确刻画性能边界，提出基于邻居边界度的增量性能边界识别方法，实现对增量数据的高效处理；为解决增量过程中的近邻搜索和反向近邻搜索的效率问题，提出改进局部敏感哈希的近似近邻搜索优化方法；分别以标准测试函数、路径规划系统作为典型的自主决策系统，开展理论研究工作的验证和分析。实验结果表明：基于邻居边界度的性能边界识别方法具有很好的方法参数泛化能力，在路径规划系统上，该方法比对比方法识别准确度高出13.68%，运行时间减少91.57%。
Abstract:The ability of an autonomous decision-making system to withstand disruptions is reflected in its performance boundary, which is a crucial indicator of its resilience. A performance boundary identification approach based on neighbor boundary degrees is proposed for autonomous decision-making systems, taking into account the features of multi-space distribution and incremental creation of performance boundary data. In order to solve the absolute scale measurement problem, we first design the neighbor boundary degree index. Then, we propose a performance boundary identification process based on neighbor boundary degree, which addresses the challenges of a complex performance boundary search space and non-uniform scale throughout the space. Secondly, the incremental performance boundary Identification method based on neighbor boundary degree is proposed by combining incremental data with the original identification results in order to accurately describe the performance boundary and to achieve efficient incremental data processing. An approximate nearest neighbor search optimization technique that enhances local sensitive hashing is then suggested in order to address the efficiency issue of nearest neighbor search and reverse nearest neighbor search that arose in the incremental phase. Finally, benchmark systems and path planning systems are used as autonomous decision-making systems to carry out verification and analysis of theoretical research work. Experimental results show that the performance boundary identification method based on neighbor boundary degree has good generalization ability of algorithm parameters. In the experiment on the path planning system, this method has a 13.68% higher boundary recognition accuracy and a 91.57% shorter running time compared to the comparison method.
-
图 1 基于邻居边界度的性能边界识别方法框图
Figure 1. Framework for performance boundary identification method based on neighbor boundary degree
图 2 路径规划系统性能边界对示例
Figure 2. Examples of path planning system performance for boundary pairs
图 4 基于邻居边界度的性能边界识别方法示意图
Figure 4. Diagram of neighbor boundary degree-based performance boundary identification method
图 5 增量数据的性能边界识别示意图
Figure 5. Illustration of identifying performance boundaries for incremental data
图 6 哈希同桶概率关系
Figure 6. Relationship between same bucket probability and Hash probability
图 11 性能边界识别方法最优F-measure与平均最优F-measure对比
Figure 11. Comparison between optimal F-measure and average parameter F-measure for performance boundary identification methods
图 12 标准测试函数中各性能边界识别方法指标对比
Figure 12. Comparison of performance boundary identification methods and indicators in standard test functions
图 13 平均最优参数下路径规划系统中各性能边界识别方法指标对比
Figure 13. Comparison of performance boundary identification methods and indicators in path planning systems under average optimal parameters
图 14 路径规划系统中性能边界识别方法最优F-measure指标对比
Figure 14. Comparison of optimal F-measure for performance boundary identification methods in path planning systems
图 15 标准测试函数中增量性能边界识别方法指标对比
Figure 15. Comparison of indicators for identifying incremental performance boundaries in standard test functions
图 16 路径规划系统中增量性能边界识别方法指标对比
Figure 16. Comparison of indicators for incremental performance boundary identification methods in path planning systems
图 17 标准测试函数中近邻搜索优化方法评价指标对比
Figure 17. Comparison of evaluation indicators for nearest neighbor search methods in standard test functions
图 18 路径规划系统中近邻搜索优化方法评价指标对比
Figure 18. Comparison of evaluation indicators for nearest neighbor search methods in path planning systems
表 1 被测系统信息
Table 1. Tested system information
标准测
试函数函数状态
空间范围函数真实
边界比例/%路径规
划系统系统状态
空间范围系统真实
边界比例/%1 [−4,4] 26.20 1 [0,500] 41.84 2 [−10,10] 76.96 2 [0,500] 51.68 3 [−50,50] 22.21 3 [0,500] 48.05 4 [−500,500] 39.80 4 [0,500] 50.68 5 [−10,10] 58.71 5 [0,500] 15.35 6 [−10,10] 40.62 6 [0,500] 20.26 7 [0,500] 20.54 8 [0,500] 17.14 
下载: 导出CSV
-
[1] HORN P. Autonomic computing: IBM’s perspective on the state of information technology[EB/OL]. (2001-3-27)[2023-11-15]. http://www.ibm.com/research/autonomic. [2] MULLINS G E, STANKIEWICZ P G, HAWTHORNE R C, et al. Adaptive generation of challenging scenarios for testing and evaluation of autonomous vehicles[J]. Journal of Systems and Software, 2018, 137: 197-215. [3] SZEGEDY C, ZAREMBA W, SUTSKEVER I, et al. Intriguing properties of neural networks[EB/OL]. (2013-12-21)[2023-11-15]. https://arxiv.org/abs/1312.6199v4. [4] GOODFELLOW I J, SHLENS J, SZEGEDY C. Explaining and harnessing adversarial examples[EB/OL]. (2015-7-16)[2023-11-15]. http://arxiv.org/abs/1412.6572. [5] SONG Y, KIM T, NOWOZIN S, et al. PixelDefend: leveraging generative models to understand and defend against adversarial examples[EB/OL]. (2018-5-23) [2023-11-15]. http://arxiv.org/abs/1710.10766. [6] ILYAS A, SANTURKAR S, TSIPRAS D, et al. Adversarial examples are not bugs, they are features[J]. Advances in Neural Information Processing Systems, 2019, 32: 132-138. [7] HEO B, LEE M, YUN S, et al. Knowledge distillation with adversarial samples supporting decision boundary[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2019, 33(1): 3771-3778. [8] PAN D Y, YIN Y J, SUN Y J, et al. Sample selection in support vector machines: a fixed neighborhood sphere approach[C]//Proceedings of the 3rd International Conference on Information Science and Control Engineering. Piscataway: IEEE Press, 2016: 1089-1093. [9] ZYKIN S, ZYKIN V. The boundary identification algorithm for a process control[C]//Proceedings of the Dynamics of Systems, Mechanisms and Machines. Piscataway: IEEE Press, 2022: 1-4. [10] CAO H Y, LI H Y, SUN W, et al. A boundary identification approach for the feasible space of structural optimization using a virtual sampling technique-based support vector machine[J]. Computers & Structures, 2023, 287: 107118. [11] ZHOU G Q, WEI H, WANG X Y, et al. BSMNet: boundary-salience multi-branch network for intima-media identification in carotid ultrasound images[J]. Computers in Biology and Medicine, 2023, 162: 107092. [12] SOWJANYA A M, SHASHI M. Cluster feature-based incremental clustering approach (CFICA) for numerical data[J]. International Journal of Computer Science and Network Security, 2010, 10: 73-79. [13] 欧阳秋宝. 基于局部分布的自组织增量聚类算法[D]. 南京: 南京大学, 2012.OUYANG Q B. Self-organizing incremental clustering algorithm based on local distribution[D]. Nanjing: Nanjing University, 2012(in Chinese). [14] ZHENG L W, HUO H, GUO Y Y, et al. Supervised adaptive incremental clustering for data stream of chunks[J]. Neurocomputing, 2017, 219: 502-517. [15] 张煜莹, 陆艺, 赵静. 基于增量学习的数控机床故障诊断系统[J]. 计量学报, 2022, 43(11): 1456-1463.ZHANG Y Y, LU Y, ZHAO J. Fault diagnosis system of numerical control machine based on incremental learning[J]. Acta Metrologica Sinica, 2022, 43(11): 1456-1463(in Chinese). [16] ESTER M, KRIEGEL H P, SANDER J, et al. Incremental clustering for mining in a data warehousing environment[C]//Proceedings of the 24th International Conference on Very Large Data Bases. New York: Morgan Kaufmann Publishers Inc, 1998: 323-333. [17] HORE P, HALL L O, GOLDGOF D B. Single pass fuzzy C means[C]//Proceedings of the IEEE International Fuzzy Systems Conference. Piscataway: IEEE Press, 2007: 1-7. [18] 夏雨轩, 曲海成, 关茜文. 一种面向图像相似度计算的改进哈希算法[J]. 智能计算机与应用, 2022, 12(4): 110-113.XIA Y X, QU H C, GUAN Q W. An improved hash algorithm for image similarity calculation[J]. Intelligent Computer and Applications, 2022, 12(4): 110-113(in Chinese). [19] DATAR M, IMMORLICA N, INDYK P, et al. Locality-sensitive hashing scheme based on p-stable distributions[C]//Proceedings of the Twentieth Annual Symposium on Computational Geometry. New York: ACM, 2004: 253-262. [20] GIONIS A, INDYK P, MOTWANI R. Similarity search in high dimensions via hashing[C]//Proceedings of the 25th International Conference on Very Large Data Bases. Edinburgh: Morgan Kaufmann, 1999: 518-529. -
下载:
点击查看大图
计量
- 文章访问数: 458
- HTML全文浏览量: 120
- PDF下载量: 50
- 被引次数: 0
