基于等价分量交叉相似性的Pareto支配性预测

郭观七; 尹呈; 曾文静; 李武; 严太山

doi:10.3724/SP.J.1004.2014.00033

基于等价分量交叉相似性的Pareto支配性预测

doi: 10.3724/SP.J.1004.2014.00033

郭观七¹,
尹呈^1,2,
曾文静^{1 3},
李武¹,
严太山¹

1.
湖南理工学院信息与通信工程学院岳阳 414006;
2.
湘潭大学信息工程学院湘潭 411105;
3.
中南大学信息科学与工程学院长沙 410083

基金项目:

国家自然科学基金（60975049）;湖南省自然科学基金（11JJ2037）;湖南省高校科技创新团队支持计划资助

计量
- 文章访问数: 1559
- HTML全文浏览量: 49
- PDF下载量: 899
- 被引次数: 0
出版历程
- 收稿日期: 2012-08-29
- 修回日期: 2013-05-02
- 刊出日期: 2014-01-20

Prediction of Pareto Dominance by Cross Similarity of Equivalent Components

1.
College of Information and Communication Engineering, Hunan Institute of Science and Technology, Yueyang 414006;
2.
Institute of Information Engineering, Xiangtan University, Xiangtan 411105;
3.
College of Information Science and Engineering, Central South University, Changsha 410083

Funds:

Supported by National Natural Science Foundation of China (60975049), Natural Science Foundation of Hunan Province (11JJ2037), and Aid Program for Science and Technology Innovative Research Team in Higher Educational Instituions of Hunan Province

摘要

摘要: 研究用最近邻分类预测多目标优化问题Pareto支配性的相似性测度方法. 在分析决策分量对各目标分量贡献率的基础上定义决策向量的等价子向量，等价子向量由贡献率相同的决策分量所组成.提出基于等价子向量的最小交叉距离加权和相似性测度方法.对每个目标分量，独立评价待测数据与N个已知样本的相似度，每个样本按其相似度值的升序赋予[0：N-1]之间的序号，按各目标上的序号之和最小准则确定最近邻样本.等价子向量最小交叉距离加权和相似性测度以及多目标最近邻搜索方法在确定决策向量相似性时，引入了决策空间到目标向量空间的映射知识，使决策变量相似性测度更真实地反映目标向量相似性.对典型多目标优化问题的Pareto支配性最近邻分类实验结果表明，提出的方法可显著地提高分类准确性.
- 多目标优化 /
- Pareto支配性 /
- 等价分量 /
- 交叉相似性 /
- 最近邻分类
Abstract: This study investigates the similarity measurement of nearest neighbor classification for Pareto dominance prediction in multi-objective optimization. The equivalent components of a decision vector are defined by analyzing the contribution rate of each decision component to each objective component. For each objective component, the decision vector is divided into a group of equivalent sub-vectors, each consisting of the equivalent components with the same contribution rate. The distance between two equivalent sub-vectors is computed by minimizing the cross distance among the equivalent components. The similarity of two decision vectors is measured by the weighted sum of the minimized cross distances (WMCDs) corresponding to each equivalent sub-vector, where the weight takes the corresponding contribution rate. For each objective component, after evaluating WMCDs between an observed data and N samples, each sample is assigned a sequential number in [0: N-1] according to the ascending order of WMCDs. The nearest sample is the one with the minimal sum of the sequential numbers on all objective components. For WMCDs and multi-objective nearest neighbor searching introduce the mapping information from the decision space to the objective space, the similarity measurement in the decision space reflects the similarity in the objective space more seemingly. The experiments on tested problems show that the proposed algorithm remarkably improves the prediction performance of nearest neighbor classifiers.
- Multi-objective optimization /
- Pareto dominance /
- equivalent components /
- cross similarity /
- nearest neighbor classiffication

HTML全文

参考文献(25)

[1]	Deb K. Multi-objective Optimisation Using Evolutionary Algorithms: An Introduction, KanGAL Report 2011003, Indian Institute of Technology Kanpur, India, 2011
[2]	Zhou A M, Qu B Y, Li H, Zhao S Z, Suganthan P N, Zhang Q F. Multiobjective evolutionary algorithms: a survey of the state of the art. Swarm and Evolutionary Computation, 2011, 1(1): 32-49
[3]	Nain P K S, Deb K. A Multi-objective Search and Optimization Procedure with Successive Approximate Models, KanGAL Report 2004012, Indian Institute of Technology Kanpur, India, 2004
[4]	Jin Y C, Sendhoff B. A systems approach to evolutionary multiobjective structural optimization and beyond. IEEE Computational Intelligence Magazine, 2009, 4(3): 62-76
[5]	Knowles J. ParEGO: A hybrid algorithm with on-line landscape approximation for expensive multiobjective optimization problems. IEEE Transactions on Evolutionary Computation, 2006, 10(1): 50-66
[6]	Schmidt M D, Lipson H. Coevolution of fitness predictors. IEEE Transactions on Evolutionary Computation, 2008, 12(6): 736-749
[7]	Samad A, Kim K Y, Goel T, Haftka R T, Shyy W. Multiple surrogate modeling for axial compressor blade shape optimization. Journal for Propulsion and Power, 2008, 24(2): 302-310
[8]	Shi L, Rasheed K. A survey of fitness approximation methods applied in evolutionary algorithms. Computational Intelligence in Expensive Optimization Problems, Adaptation Learning and Optimization. Berlin Heidelberg: Springer, 2010. 2: 3-28
[9]	Jin Y C. Surrogate-assisted evolutionary computation: recent advances and future challenges. Swarm and Evolutionary Computation, 2011, 1(2): 61-70
[10]	Goel T, Vaidyanathan R, Haftka R T, Shyy W, Queipo N V, Tucker K. Response surface approximation of Pareto opimal front in multi-objective optimization. Computer Methods in Applied Mechanics and Engineering, 2007, 196(4-6): 879-893
[11]	Zhang Q F, Liu W D, Tsang E, Virginas B. Expensive multiobjective optimization by MOEA/D with Gaussian process model. IEEE Transactions on Evolutionary Computation, 2010, 14(3): 456-474
[12]	Deb K, Nain P K S. An evolutionary multi-objective adaptive meta-modeling procedure using artificial neural networks. Evolutionary Computation in Dynamic and Uncertain Environments. Berlin, Germany: Springer-Verlag, 2007. 51: 297-322
[13]	Marjavaara B D, Lundström T S, Goel T, Mack Y, Shyy W. Hydraulic turbine diffuser shape optimization by multiple surrogate model approximations of Pareto fronts. Journal of Fluids Engineering, 2007, 129(9): 1228-1240
[14]	Lim D, Ong Y S, Jin Y C, Sendhoff B. A study on metamodeling techniques, ensembles, and multi-surrogates in evolutionary computation. In: Proceedings of the 9th Annual Conference on Genetic and Evolutionary Computation. New York, USA: ACM, 2007. 1288-1295
[15]	Lim D, Jin Y C, Ong Y S, Sendoff B. Generalizing surrogate-assisted evolutionary computation. IEEE Transactions on Evolutionary Computation, 2010, 14(3): 329-355
[16]	Acar E, Rais-Rohani M. Ensemble of metamodels with optimized weight factors. Structural and Multidisciplinary Optimization, 2009, 37(3): 279-294
[17]	Shi L, Rasheed K. ASAGA: an adaptive surrogate-assisted genetic algorithm. In: Proceedings of the 10th Annual Conference on Genetic and Evolutionary Computation. New York, USA: ACM, 2008. 1049-1056
[18]	Guo G, Li W, Yang B, Li W, Yin C. Predicting Pareto dominance in multi-objective optimization using pattern recognition. In: Proceedings of the 2012 International Conference on Intelligent System Design and Engineering Application (ISDEA). Sanya, China: IEEE, 2012. 456-459
[19]	Schaffer J D. Multiple objective optimization with vector evaluated genetic algorithms. In: Proceedings of the 1st International Conference on Genetic Algorithms. New Jersey, USA: Lawrence Erlbaum, 1987. 93-100
[20]	Guo G Q, Yin C, Yan T S, Li W B. Binary nearest neighbor classification of predicting Pareto dominance in multi-objective optimization. In: Proceedings of the 2012 International Conference on Swarm Intelligence, Part I, LNCS 7331. Berlin: Springer-Verlag, 2012. 537-545
[21]	Guo G Q, Yin C, Yan T S, Li W. Nearest neighbor classification of Pareto dominance in multi-objective optimization. In: Proceedings of the 5th IEEE International Conference on Advanced Computational Intelligence. Nanjing, China: IEEE, 2012. 328-331
[22]	Fonseca C M, Fleming P J. Multiobjective optimization and multiple constraint handling with evolutionary algorithm Part Ⅱ: Application example. IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans, 1998, 28(1): 38-47
[23]	Kursawe F. A variant of evolution strategies for vector optimization. Parallel Problem Solving from Nature, Schwefel I H P, Manner R (eds.), Berlin, Germany: Springer, 1990. 193-197
[24]	Zilter E, Deb K, Thiele L. Comparison of multiobjective evolutionary algorithms: empirical results. Evolutionary Computation, 2000, 8(2): 173-195
[25]	Deb K, Thiele L, Laumanns M, Zitzler E. Scalable multi-objective optimization test problems. In: Proceedings of the 2002 Congress on Evolutionary Computation (CEC'02). Honolulu, HI: IEEE, 2002. 825-830