基于BP神经网络的中部地区物流需求预测
|
摘要
中部地区位于我国的核心地带,在我国的经济社会发展中具有非常重要的地位。国家领导也多次在报告中提出要“促进中部地区崛起”,把中部地区建成为我国重要的物流、商流、信息流集散地。中部地区有着较为优越的地理区位优势,但发展上仍然受到经济条件的制约。目前物流建设的速度正在稳步提升,但是交通网络及物流服务水平仍与发达国家及东部发达地区存在较大的差距。快速有效的发展中部地区的物流产业,对促进中部地区的崛起有着至关重要的作用。
本文将采用理论与实证相结合的方法,分析中部地区的物流现状,研究影响中部地区物流需求的主要因素,选择选择人工神经网络作为预测方法,对中部地区的物流规模需求和结构需求进行合理预测,并分析相关预测数据的合理性和有效性,为中部地区未来一定时期内的物流系统规划提供理论依据。
本文首先对物流需求预测进行理论研究,通过对比来分析有效的物流需求预测指标,介绍神经网络相关理论,总结BP神经网络的不足之处和改进方法,说明将神经网络应用于物流需求预测的可行性和优越性;其次进行中部地区物流需求预测的指标体系设计,研究中部地区的经济和物流发展现状,分析区域物流需求的量化指标,选择具有代表作用的指标作为我们的预测指标,收集历史数据,进行相关性分析,构建相应的预测指标体系;再次,根据物流规模需求和物流结构需求预测的不同指标体系,建立不同的神经网络,根据构建的预测网络预测未来三年的物流需求,将传统的需求预测方法与通过神经网络预测得到的数据进行对比分析,确定神经网络预测方法的优越性和准确性,分析BP神经网络的改进算法,进行算法的调整分析,建立改进的神经网络后对今后三年的中部地区货运量进行需求预测,并分析结果。
The central region as the belly heart of the whole country, it is a quite important position to occupy in national development of social and economic. The leaders of the Central Party Committee, propose“promoting the central region to emerge”in the report more than once, which make great efforts to build central region into the national important granary, important base of energy and raw materials, national important distributing centre of logistics, trade and information flow. The construction speed of the logistics infrastructure in the central region is being accelerated. Now it has already formed the preliminary and perfect comprehensive traffic network, but compared with developed country and developed area of the east region of our country still has greater disparity. Effective development of central region’s logistics industry is quite important for the promotion of the emergence and utilize of geography and environmental resources of the central region.
This text will adopt the method which combines with theory and real example, will analyze the logistics current situation of the central region, and research the main factor which influence the logistics demand in the central region. Then we choose the artificial neural network as the method to predict the demand of scale and structure of logistics demand in the middle part rationally. Then we analyze the accuracy and effectiveness of the predicted data which can offer the theoretical foundation for the logistics system plan in the future of central region.
This text carries on the theoretical research to the requirement forecasting of the logistics at first, choose the effective indicators for logistics requirement forecasting, then introduce the relevant theory of the neural network, summarize the weakness and improvement method of BP neural network, prove the feasibility and superiority for applying the neural network to the requirement forecasting of logistics. Second, we design the index system for logistics requirement forecasting of central region, then study the current situation of the development of logistics and economy and analyze the quantization index of the regional logistics demand, choose the representative indicators. We can collect the historical data, analyze the relevance and structure the corresponding prediction indicators system. Third, we can build different neural networks according to different indicators system of logistics scale demand and structure requirement forecasting. Moreover, we can predict the following three years according and compare the data result using the traditional requirement forecasting method with the neural network which can confirm that the neural network can predict the superiority and accuracy of the method. Later, we should analyze the improvement algorithm of BP neural network and use the adjusted algorithm to forecast the volume of goods transported of central region of the following three years.
本文将采用理论与实证相结合的方法,分析中部地区的物流现状,研究影响中部地区物流需求的主要因素,选择选择人工神经网络作为预测方法,对中部地区的物流规模需求和结构需求进行合理预测,并分析相关预测数据的合理性和有效性,为中部地区未来一定时期内的物流系统规划提供理论依据。
本文首先对物流需求预测进行理论研究,通过对比来分析有效的物流需求预测指标,介绍神经网络相关理论,总结BP神经网络的不足之处和改进方法,说明将神经网络应用于物流需求预测的可行性和优越性;其次进行中部地区物流需求预测的指标体系设计,研究中部地区的经济和物流发展现状,分析区域物流需求的量化指标,选择具有代表作用的指标作为我们的预测指标,收集历史数据,进行相关性分析,构建相应的预测指标体系;再次,根据物流规模需求和物流结构需求预测的不同指标体系,建立不同的神经网络,根据构建的预测网络预测未来三年的物流需求,将传统的需求预测方法与通过神经网络预测得到的数据进行对比分析,确定神经网络预测方法的优越性和准确性,分析BP神经网络的改进算法,进行算法的调整分析,建立改进的神经网络后对今后三年的中部地区货运量进行需求预测,并分析结果。
The central region as the belly heart of the whole country, it is a quite important position to occupy in national development of social and economic. The leaders of the Central Party Committee, propose“promoting the central region to emerge”in the report more than once, which make great efforts to build central region into the national important granary, important base of energy and raw materials, national important distributing centre of logistics, trade and information flow. The construction speed of the logistics infrastructure in the central region is being accelerated. Now it has already formed the preliminary and perfect comprehensive traffic network, but compared with developed country and developed area of the east region of our country still has greater disparity. Effective development of central region’s logistics industry is quite important for the promotion of the emergence and utilize of geography and environmental resources of the central region.
This text will adopt the method which combines with theory and real example, will analyze the logistics current situation of the central region, and research the main factor which influence the logistics demand in the central region. Then we choose the artificial neural network as the method to predict the demand of scale and structure of logistics demand in the middle part rationally. Then we analyze the accuracy and effectiveness of the predicted data which can offer the theoretical foundation for the logistics system plan in the future of central region.
This text carries on the theoretical research to the requirement forecasting of the logistics at first, choose the effective indicators for logistics requirement forecasting, then introduce the relevant theory of the neural network, summarize the weakness and improvement method of BP neural network, prove the feasibility and superiority for applying the neural network to the requirement forecasting of logistics. Second, we design the index system for logistics requirement forecasting of central region, then study the current situation of the development of logistics and economy and analyze the quantization index of the regional logistics demand, choose the representative indicators. We can collect the historical data, analyze the relevance and structure the corresponding prediction indicators system. Third, we can build different neural networks according to different indicators system of logistics scale demand and structure requirement forecasting. Moreover, we can predict the following three years according and compare the data result using the traditional requirement forecasting method with the neural network which can confirm that the neural network can predict the superiority and accuracy of the method. Later, we should analyze the improvement algorithm of BP neural network and use the adjusted algorithm to forecast the volume of goods transported of central region of the following three years.
引文
[1]吴清一.物流管理[M].中国物资出版社,2005
[2]龚树生.我国物流统计发展现状分析[J].中国流通经济. 2005,24(9):23-28
[3]佘廉,张晓燕.我国中部五省物流网络研究[J].物流与交通,2005, 36(7):33-35
[4]《政府工作报告》温家宝,2004
[5]黄军根.中部地区物流产业与区域经济联系研究[D].南昌大学. 2007:33-36
[6] Martion Christopher. Logistic and supply chain management[J]. London: Pitmin publishing 2002,20(8):25-29
[7] Patric T. Harker, Predicting Intercity Freight Flows[M], VNU Science Press BV,1997
[8] Spay, R.H., &Friedlaender A.F., Hedonic cost functions for the regulated trucking industry[J], The Bell Journal of Economics, 1997,12(9),: 159~179
[9] Vivien P Jeffs & Peter J Hills, Determinants of model choice in freight transport: A case study[J], 200024(17),29~47
[10] T. Hill, L, Marquez, etc., Artificial Neural Network Model for Forecasting and Decision-making[J], International Journal of Forecasting, 2004,36(10):4-15
[11]王晓忠,物流量预测方法研究[D].武汉理工大学硕士论文,2005:42-46
[12] S. Makridakis, A. Andersen, R. Carbone, etc, The accuracy of Extrapolation Methods[J]: Results of A Forecasting Competition. Journal of Forecasting, 2002,20(12): 111-152
[13]程肖冰,张群.区域物流需求预测方法比较分析[J].工业工程与管理. 2008,12(6): 95-99
[14] Bates J.N, Granger C.W.J Combination of forecasts[J]. Operations Research Quarterly, 1999,20(4):451-468
[15] Rumelhartde, Hinton G E,Williams R J. Learing internal representations by error propagation. Parallel Data Processing[J], Cambridge, MA: The M.I.T.Press,1998,36(10):318-362
[16]陈曦,庞叶.一种新的集成预测方法——GPVECM[J].系统工程理论与实践,2008,20(12):108-112
[17]初良勇,田质广,谢新连.组合预测模型在物流需求预测中的应用[J].物流经济. 2004,12(4):43-46
[18] Moller M F. A scaled conjugate gradient algorithm for fast supervised learning. Neural Networks[J],2003,12(6):525-533
[19] Friesz, T,L. ,et al, The Northeast Regional environmental Impact Study: Theory, Validation and Application of a Freight Network Equilibrium Model, Presented at the Transportation Equilibrium and Supply Model[J], Argonne, IL. 1991,20(11):155-160
[20] Dutta, Decision Support in Non-conservative Domains: Generalization with Neural Networks[J], Decision Support Systems, 1994,20(11): 527-544
[21]杨茂盛,孙珂.人工神经网络技术在物流需求量预测中的应用[J].物流技术. 2005,12(5): 39-41
[22]秦立公,张建,杨一俊.基于人工神经网络的时间序列分析方法在物流需求量预测中的应用[J].物流科技. 2007,12(1): 03-05
[23]林荣天,陈联诚,李绍静.基于灰色神经网络的区域物流需求预测.价值工程. 2007,12(2): 92-95
[24] Hagan M T, Menhaj M. Training feedforward networks with the Marquard algorithm IEEE Transactions on Neural Networks[J], 2004,5(6):989-993
[25]贾星辰,王铁宁,裴帅.基于BP神经网络的物流需求量预测模型研究. [J].物流科技.2007,60(29): 03-05
[26]张凤荣,金俊武,李延忠.基于改进的灰色BP神经网络的区域物流成本预测[J].公路交通科技. 2005年6月第22卷155-158
[27] Nguyen D, Widrow B. Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights. Proceedings of the International Joint Conference on Neural Networks[J]. Washingtonm. 1999,20(3):21-26
[28] Lancaster, K.J., A new approach to consumer thery[J], J. of Pol.Econ.7, 1997, 20(17):132~157
[29] ECMT, Goods distribution System in Urban Areas, Report of the Sixty First Round Table on Transport Economics[M],2004
[30]王红玉,李琰.中国物流业发展现状及其对策研究[J].物流经济.2006,24(12):63-65
[31] Joseph Sussman. Introduction to transportation system. US[M]: Norwood Artech house. 2000
[32] James R.Stock, Douglas M.Lambert. Strategic Logistics Management[M]. McGraw-Hill publishing firm, 2001:5
[33]李志.区域物流产业构成及其基本特征.中国流通经济. 2008,36(4): 15-18
[34] Williams R J. A learning algorithm for continually running fully recurrent NN[J]. Neural Computation, 1989,12(1): 270-280.
[35] Francl, L. S. P. Artificial neural network models of wheat leaf wetness [J].1997,20(5):57-65
[36] Ray, R., Hashemi, L. A. A Neural Netwok for Transportation satety Modeling[J].1995,12(9): 247-256
[37]韩立群.人工神经网络理论、设计及应用[M].化学工业出版社,2007
[38] Joseph Sussman. Introduction to transportation system. US[M]: Norwood Artech house. 2000
[39] PingSun Leung, L. T. T. Predicting shrimp disease occurrence: artifificial neural networks vs logistic regression[J].1999,12(5):15-24
[40]王静伟. BP神经网络改进算法的研究[J].计算机技术与发展. 2008,12(8): 157-158
[41] Shen Miqun, Shang Guoqing, Tong Dechun. Study on a technology of the to-and-fro mechanism state inspection and fault identification[J]. Journal of Vibration, Measurement & Diagnosis,1996,16(2):7-12
[42]王俊清. BP神经网络及其改进[J].重庆工学院学报(自然科学版) 2007.20(3) :75-78
[43]胡耀垓,李凯扬,钟毓宁.一种改进的神经网络BP算法[J].武汉大学学报(自然科学版).2004.45(1):25-29
[44] Nagy, H. M. , Watanabe, K. Prediction of Sediment Load Concentration in Riversusing[M].2005
[45] Karpenko M. A Neural network based fault detection and identification scheme for pneumatic process control valves[J]. IEEE International conference on systems, Man, and Cybernetics. 2001,36(17): 93-98
[46]魏然.我国中部地区物流发展现状及问题分析[J].物流技术2008,27(9):75-79
[47]张文杰.区域经济发展与现代物流[J].中国流通经济2004,12(7): 66-69
[48]余彩艳.中部地区物流需求的实证分析[D].南昌大学硕士论文2008
[49] Allen, W. Bruce. The demand for freight transportation, A micro approach[J], Transpon. 1997,12(11):205-211
[50] Joseph Sussman. Introduction to transportation systems[J]. US: Norwood Artech house. 2000,20(9): 108-117
[51]徐杰.区域经济的发展对区域物流需求的影响[J].数量经济研究. 2003,12(7): 24-29
[52] Barro R and Sala Martin. Economic Growth[M], New York: Mc Graw- Book Company, 2005
[53]区域物流业对地区经济增长的影响分析[J].统计与决策,2006,12(2): 109-112
[54]谢如鹤.物流系统规划原理与方法[M].中国物资出版社,2004
[2]龚树生.我国物流统计发展现状分析[J].中国流通经济. 2005,24(9):23-28
[3]佘廉,张晓燕.我国中部五省物流网络研究[J].物流与交通,2005, 36(7):33-35
[4]《政府工作报告》温家宝,2004
[5]黄军根.中部地区物流产业与区域经济联系研究[D].南昌大学. 2007:33-36
[6] Martion Christopher. Logistic and supply chain management[J]. London: Pitmin publishing 2002,20(8):25-29
[7] Patric T. Harker, Predicting Intercity Freight Flows[M], VNU Science Press BV,1997
[8] Spay, R.H., &Friedlaender A.F., Hedonic cost functions for the regulated trucking industry[J], The Bell Journal of Economics, 1997,12(9),: 159~179
[9] Vivien P Jeffs & Peter J Hills, Determinants of model choice in freight transport: A case study[J], 200024(17),29~47
[10] T. Hill, L, Marquez, etc., Artificial Neural Network Model for Forecasting and Decision-making[J], International Journal of Forecasting, 2004,36(10):4-15
[11]王晓忠,物流量预测方法研究[D].武汉理工大学硕士论文,2005:42-46
[12] S. Makridakis, A. Andersen, R. Carbone, etc, The accuracy of Extrapolation Methods[J]: Results of A Forecasting Competition. Journal of Forecasting, 2002,20(12): 111-152
[13]程肖冰,张群.区域物流需求预测方法比较分析[J].工业工程与管理. 2008,12(6): 95-99
[14] Bates J.N, Granger C.W.J Combination of forecasts[J]. Operations Research Quarterly, 1999,20(4):451-468
[15] Rumelhartde, Hinton G E,Williams R J. Learing internal representations by error propagation. Parallel Data Processing[J], Cambridge, MA: The M.I.T.Press,1998,36(10):318-362
[16]陈曦,庞叶.一种新的集成预测方法——GPVECM[J].系统工程理论与实践,2008,20(12):108-112
[17]初良勇,田质广,谢新连.组合预测模型在物流需求预测中的应用[J].物流经济. 2004,12(4):43-46
[18] Moller M F. A scaled conjugate gradient algorithm for fast supervised learning. Neural Networks[J],2003,12(6):525-533
[19] Friesz, T,L. ,et al, The Northeast Regional environmental Impact Study: Theory, Validation and Application of a Freight Network Equilibrium Model, Presented at the Transportation Equilibrium and Supply Model[J], Argonne, IL. 1991,20(11):155-160
[20] Dutta, Decision Support in Non-conservative Domains: Generalization with Neural Networks[J], Decision Support Systems, 1994,20(11): 527-544
[21]杨茂盛,孙珂.人工神经网络技术在物流需求量预测中的应用[J].物流技术. 2005,12(5): 39-41
[22]秦立公,张建,杨一俊.基于人工神经网络的时间序列分析方法在物流需求量预测中的应用[J].物流科技. 2007,12(1): 03-05
[23]林荣天,陈联诚,李绍静.基于灰色神经网络的区域物流需求预测.价值工程. 2007,12(2): 92-95
[24] Hagan M T, Menhaj M. Training feedforward networks with the Marquard algorithm IEEE Transactions on Neural Networks[J], 2004,5(6):989-993
[25]贾星辰,王铁宁,裴帅.基于BP神经网络的物流需求量预测模型研究. [J].物流科技.2007,60(29): 03-05
[26]张凤荣,金俊武,李延忠.基于改进的灰色BP神经网络的区域物流成本预测[J].公路交通科技. 2005年6月第22卷155-158
[27] Nguyen D, Widrow B. Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights. Proceedings of the International Joint Conference on Neural Networks[J]. Washingtonm. 1999,20(3):21-26
[28] Lancaster, K.J., A new approach to consumer thery[J], J. of Pol.Econ.7, 1997, 20(17):132~157
[29] ECMT, Goods distribution System in Urban Areas, Report of the Sixty First Round Table on Transport Economics[M],2004
[30]王红玉,李琰.中国物流业发展现状及其对策研究[J].物流经济.2006,24(12):63-65
[31] Joseph Sussman. Introduction to transportation system. US[M]: Norwood Artech house. 2000
[32] James R.Stock, Douglas M.Lambert. Strategic Logistics Management[M]. McGraw-Hill publishing firm, 2001:5
[33]李志.区域物流产业构成及其基本特征.中国流通经济. 2008,36(4): 15-18
[34] Williams R J. A learning algorithm for continually running fully recurrent NN[J]. Neural Computation, 1989,12(1): 270-280.
[35] Francl, L. S. P. Artificial neural network models of wheat leaf wetness [J].1997,20(5):57-65
[36] Ray, R., Hashemi, L. A. A Neural Netwok for Transportation satety Modeling[J].1995,12(9): 247-256
[37]韩立群.人工神经网络理论、设计及应用[M].化学工业出版社,2007
[38] Joseph Sussman. Introduction to transportation system. US[M]: Norwood Artech house. 2000
[39] PingSun Leung, L. T. T. Predicting shrimp disease occurrence: artifificial neural networks vs logistic regression[J].1999,12(5):15-24
[40]王静伟. BP神经网络改进算法的研究[J].计算机技术与发展. 2008,12(8): 157-158
[41] Shen Miqun, Shang Guoqing, Tong Dechun. Study on a technology of the to-and-fro mechanism state inspection and fault identification[J]. Journal of Vibration, Measurement & Diagnosis,1996,16(2):7-12
[42]王俊清. BP神经网络及其改进[J].重庆工学院学报(自然科学版) 2007.20(3) :75-78
[43]胡耀垓,李凯扬,钟毓宁.一种改进的神经网络BP算法[J].武汉大学学报(自然科学版).2004.45(1):25-29
[44] Nagy, H. M. , Watanabe, K. Prediction of Sediment Load Concentration in Riversusing[M].2005
[45] Karpenko M. A Neural network based fault detection and identification scheme for pneumatic process control valves[J]. IEEE International conference on systems, Man, and Cybernetics. 2001,36(17): 93-98
[46]魏然.我国中部地区物流发展现状及问题分析[J].物流技术2008,27(9):75-79
[47]张文杰.区域经济发展与现代物流[J].中国流通经济2004,12(7): 66-69
[48]余彩艳.中部地区物流需求的实证分析[D].南昌大学硕士论文2008
[49] Allen, W. Bruce. The demand for freight transportation, A micro approach[J], Transpon. 1997,12(11):205-211
[50] Joseph Sussman. Introduction to transportation systems[J]. US: Norwood Artech house. 2000,20(9): 108-117
[51]徐杰.区域经济的发展对区域物流需求的影响[J].数量经济研究. 2003,12(7): 24-29
[52] Barro R and Sala Martin. Economic Growth[M], New York: Mc Graw- Book Company, 2005
[53]区域物流业对地区经济增长的影响分析[J].统计与决策,2006,12(2): 109-112
[54]谢如鹤.物流系统规划原理与方法[M].中国物资出版社,2004