一种基于神经网络的仿真优化方法

doi:10.16182/j.issn1004731x.joss.201801005

摘要/Abstract

摘要： 为提高仿真优化问题求解效率,提出了一种基于神经网络的仿真优化方法。利用神经网络对非线性输入输出关系的逼近能力,由神经网络输出值代替仿真结果以减少所需仿真次数。按照提出的3种样本选择策略,由仿真模型产生一定数量的样本,借助广义回归神经网络在学习速度、网络稳定性、参数选取方面的独特优势,对样本进行训练,生成能够反映仿真模型输入输出关系的回归曲面,以实现用神经网络输出值代替仿真结果,利用优化算法对回归曲面进行寻优。通过对典型测试函数进行实验,证明了方法的可行性和有效性。

关键词: 神经网络, 仿真优化, 回归曲面, 样本选择

Abstract: To improve the efficiency of optimization via simulation (OvS), an OvS method based on neural network is proposed. Taking advantage of the approximation ability of neural network to nonlinear input-output relationship, neural network's outputs are used as substitutes for simulation results to reduce the required simulation runs. Samples are generated by simulation according to the three proposed samples selection methods. Owning to its advantages on learning speed, network stability and parameters selection, generalized regression neural network (GRNN) is adopted to train the samples. The trained GRNN forms a regression surface that represents the relationship between simulation inputs and outputs, which makes it feasible to use GRNN output as substitutes for simulation runs. Optimization algorithms are applied to search for the best solution on the regression surface. Experiments are carried out with some typical test functions, and the feasibility and effectiveness of our method are demonstrated.

Key words: neural network, optimization via simulation, regression surface, samples selection

中图分类号:

TP391.9

吴诗辉, 刘晓东, 邵悦, 张发, 杨闽湘. 一种基于神经网络的仿真优化方法[J]. 系统仿真学报, 2018, 30(1): 36-44.

Wu Shihui, Liu Xiaodong, Shao Yue, Zhang Fa, Yang Minxiang. Optimization via Simulation Based on Neural Network[J]. Journal of System Simulation, 2018, 30(1): 36-44.

参考文献

[1] 孙剑, 刘好德, 李克平. 城市干道交通信号协调控制仿真优化[J]. 同济大学学报(自然科学版), 2009, 37(11): 1467-1471.Sun Jian, Liu Haode, Li Keping. Simulation optimization for urban arterial coordinate control [J]. Journal of Tongji University (Natural Science), 2009, 37(11): 1467-1471.
[2] 李东, 汪定伟. 基于仿真的铁路集装箱装卸策略优化[J]. 系统仿真学报, 2009, 21(11): 3188-3191. Li Dong, Wang Dingwei. Simulation-based optimization of loading-unloading strategies in railway container terminal[J]. Journal of System Simulation, 2009, 21(11): 3188-3191.
[3] Tsai S C, Zheng Y X.A simulation optimization approach for a two-echelon inventory system with service level constraints[J]. European Journal of Operational Research (S0377-2217), 2013, 229(2): 364-374.
[4] 刘娟, 张伟, 李群, 等. 基于启发式仿真优化的航迹规划方法研究[J]. 系统仿真学报, 2011, 23(12): 2592-2596. Liu Juan, Zhang Wei, Li Qun, et al. Research of heuristic simulation optimization-based path planning[J]. Journal of System Simulation, 2011, 23(12): 2592-2596.
[5] Hong L J, Nelson B L.A brief introduction to optimization via simulation[C]// Proceedings of the 2009 Winter Simulation Conference. Piscataway, New Jersey, USA: IEEE, 2009: 75-85.
[6] 张会奇, 陈春良, 刘峻岩, 等. 装甲装备维修保障资源优化仿真研究[J]. 系统仿真学报, 2015, 27(1): 142-146. Zhang Huiqi, Chen Chunliang, Liu Junyan, et al. Simulation research on armored equipment maintenance support resource optimization [J]. Journal of System Simulation, 2015, 27(1): 142-146.
[7] Wang L F, Shi L Y.Simulation optimization: a review on theory and applications[J]. Acta Automatica Sinica (S1874-1029), 2013, 39(11): 1957-1968.
[8] Hong L J, Nelson B L.Discrete optimization via simulation using COMPASS[J]. Operations Research (S0030-364X), 2006, 54(1): 115-129.
[9] Kabirian A, Ólafsson S.Continuous optimization via simulation using Golden Region search[J]. European Journal of Operational Research (S0377-2217), 2011, 208(1): 19-27.
[10] Khoo L P, Chen C H.Integration of response surface methodology with genetic algorithms[J]. Int J Advanced Manufacturing Technology (S0268-3768), 2001, 18(7): 483-489.
[11] Capocchi L, Santucci J F.Discrete optimization via simulation of catchment basin management within the devsimpy framework[C]//Proceedings of the 2013 Winter Simulation Conference. Piscataway, New Jersey, USA: IEEE, 2013: 205-216.
[12] Specht D F.A general regression neural network[J]. IEEE Trans Neural Networks(S1045-9227), 1991, (2): 568-576.
[13] Tomandl D, Schober A.A modified general regression neural network (MGRNN) with new, efficient training algorithms as a robust 'black box'-tool for data analysis[J]. Neural Netw(S0893-6080), 2001, 14(8): 1023-1034.
[14] Li Q, Meng Q, Cai J, et al.Predicting hourly cooling load in the building: A comparison of support vector machine and different artificial neural networks[J]. Energy Conversion and Management (S0196-8904), 2009(50): 90-96.
[15] Ward Systems Group, Inc. Neuroshell/Neurowindows Release 3.0 Manual, 3rd Edition[M]. Maryland, 1996.
[16] 景涛. 基于改进广义回归神经网络的雷达故障预测[J]. 科学技术与工程, 2009, 9(15): 4492-4494. Jing Tao. Fault prediction of radar based on modified general regression neural network[J]. Science Technology and Engineering, 2009, 9(15): 4492-4494.
[17] Keka S, Papita D S, Siddhartha D.Response surface optimization and artificial neural network modeling of microwave assisted natural dye extraction from pomegranate rind[J]. Industrial Crops and Products (S0926-6690), 2012, (37): 408-414.
[18] Li C F, Bovik A C, Wu X J.Preprocessing data for short-term load forecasting with a general regression neural network and a moving average filter[J]. IEEE Trans Neural Networks (S1045-9227), 2011, 22(5): 793-799.
[19] 薛定宇, 陈阳泉. 高等应用数学问题的MATLAB求解 [M]. 3版. 北京: 清华大学出版社, 2013. Xue D Y, Chen Y Q. Advanced applied mathematical problem solutions with Matlab [M]. 3rd edition. Beijing: Tsinghua University Press, 2013.