本文選題：灰色神經(jīng)網(wǎng)絡 + 港口物流需求預測��； 參考：《天津大學》2012年碩士論文

【摘要】：隨著近些年中國外貿(mào)交易額的持續(xù)高速增長以及世界經(jīng)濟一體化程度的加深，，世界各國的大型港口特別是樞紐港將在推進經(jīng)濟發(fā)展的歷程中發(fā)揮重要作用，尤其對于港口資源及腹地企業(yè)資源的配置方面。隨著全球生產(chǎn)與制造業(yè)正在逐步向亞洲尤其是中國轉(zhuǎn)移，我國沿海港口將承擔著重要的貨物疏散任務，所以加快發(fā)展港口現(xiàn)代物流業(yè)是經(jīng)濟發(fā)展的客觀要求，也是我國港口物流業(yè)做強做大的重要機遇，抓住這一機遇的前提是對港口物流需求的發(fā)展趨勢進行預判，尤其是港口物貨物和集裝箱吞吐量水平的變化。 考慮到物流需求的非線性變化特點及我國物流數(shù)據(jù)統(tǒng)計不完善的特殊情況，本文創(chuàng)造性地將灰色理論與神經(jīng)網(wǎng)絡算法相結(jié)合，以克服數(shù)據(jù)貧乏和數(shù)據(jù)非線性的困難。所以，作者首先對灰色系統(tǒng)理論和人工神經(jīng)網(wǎng)絡理論進行簡述，通過分析影響區(qū)域物流需求的各項指標來研究港口物流需求發(fā)展問題，并進一步以天津港為研究對象進行實證分析。作者簡單介紹物流需求預測的基本理論和方法，包括區(qū)域物流需求預測指標選取、灰色理論、神經(jīng)網(wǎng)絡算法等，這是后續(xù)實證工作展開的理論基礎。其次，重點分析了影響港口物流需求的五方面因素，即經(jīng)濟水平、產(chǎn)業(yè)結(jié)構(gòu)、消費水平、區(qū)域貿(mào)易及固定投資額，通過對五方面影響因素的分析，提取出用于天津港港口實際需求預測的二級指標集。然后，根據(jù)港口物流需求預測指標集的設定以及對算法可行性分析的基礎上，構(gòu)建了用于港口物流需求預測的灰色神經(jīng)網(wǎng)絡組合模型，該模型以影響指標和時間因素共九個指標作為輸入，以港口貨物吞吐量作為輸出，實證研究表明模型對輸入與輸出之間的非線性關系進行了較好的擬合。最后，文章以天津港港口物流需求預測為實證研究的對象，對港口未來幾年的發(fā)展給出預測結(jié)果，并根據(jù)預測結(jié)果及國內(nèi)外成熟港口的成長模式提出天津港港口物流發(fā)展的五大方面策略，包括陸上疏運結(jié)構(gòu)、內(nèi)河運輸、港口物流功能、港企合作和信息平臺建設等。 研究結(jié)果表明，采用灰色預測理論與非線性預測功能的神經(jīng)網(wǎng)絡的組合算法，能夠有效地發(fā)現(xiàn)港口物流需求影響因素與輸出指標之間的聯(lián)系，本文的實證研究有效驗證了該算法的可靠性和可行性，為研究港口物流需求預測乃至區(qū)域物流需求預測提供了另一種思路。
[Abstract]:With the rapid growth of China's foreign trade volume and the deepening of the integration of the world economy in recent years, the large ports around the world, especially the hub ports, will play an important role in the process of promoting economic development. Especially for the allocation of port resources and hinterland enterprise resources. With the gradual transfer of global production and manufacturing to Asia, especially China, China's coastal ports will undertake the important task of cargo evacuation, so speeding up the development of modern port logistics is the objective requirement of economic development. It is also an important opportunity for China's port logistics industry to become stronger and bigger. The premise of seizing this opportunity is to pre-judge the development trend of port logistics demand, especially the change of port cargo and container throughput level. Considering the characteristics of nonlinear change of logistics demand and the special situation of incomplete logistics data statistics in China, this paper creatively combines grey theory with neural network algorithm to overcome the difficulties of data scarcity and data nonlinearity. Therefore, firstly, the author makes a brief introduction to the grey system theory and artificial neural network theory, and studies the port logistics demand development by analyzing the indexes that affect the regional logistics demand. And further take Tianjin Port as the research object to carry on the empirical analysis. The author briefly introduces the basic theories and methods of logistics demand forecasting, including the selection of regional logistics demand forecasting indicators, grey theory, neural network algorithm, etc. This is the theoretical basis of the subsequent empirical work. Secondly, the paper analyzes the five factors that affect port logistics demand, that is, economic level, industrial structure, consumption level, regional trade and fixed investment. The second level index set for actual demand forecast of Tianjin Port is extracted. Then, according to the setting of port logistics demand forecasting index set and the feasibility analysis of the algorithm, a grey neural network combination model for port logistics demand forecasting is constructed. The model takes nine indexes of influence index and time factor as input and port cargo throughput as output. The empirical research shows that the nonlinear relationship between input and output is well fitted by the model. Finally, the article takes Tianjin port logistics demand forecast as the empirical research object, gives the forecast result to the port development in the next few years. According to the forecast results and the growth model of domestic and foreign mature ports, the paper puts forward five major strategies of port logistics development of Tianjin Port, including land transport structure, inland river transport, port logistics function, cooperation between Hong Kong and enterprises and information platform construction. The research results show that the combination algorithm of grey forecasting theory and nonlinear forecasting function can effectively find the relationship between port logistics demand influencing factors and output indexes. The empirical research in this paper effectively verifies the reliability and feasibility of the algorithm, and provides another way of thinking for the study of port logistics demand prediction and regional logistics demand prediction.
【學位授予單位】：天津大學
【學位級別】：碩士
【學位授予年份】：2012
【分類號】：F259.23;F552.6

【參考文獻】

相關期刊論文 前10條

1 何國華;;區(qū)域物流需求預測及灰色預測模型的應用[J];北京交通大學學報(社會科學版);2008年01期

2 曹萍;陳福集;;GA-灰色神經(jīng)網(wǎng)絡的區(qū)域物流需求預測[J];北京理工大學學報(社會科學版);2012年01期

3 王玲;劉育龍;;基于主成分分析和全回歸法的天津港貨物吞吐量的預測[J];港口經(jīng)濟;2010年04期

4 李斌,許仕榮,柏光明,李黎武;灰色—神經(jīng)網(wǎng)絡組合模型預測城市用水量[J];中國給水排水;2002年02期

5 尚鋼,鐘珞,閆京生;兩種灰色神經(jīng)網(wǎng)絡模型及應用[J];武漢理工大學學報;2002年12期

6 劉源;;基于灰色預測模型的物流需求分析[J];物流技術;2012年11期

7 馬文君;李靜;;基于灰色GM(1,1)模型的河北沿海地區(qū)物流需求預測研究[J];物流技術;2012年11期

8 高菲菲;;基于灰色BP神經(jīng)網(wǎng)絡的汽車物流需求量預測模型[J];中國新技術新產(chǎn)品;2011年18期

9 王明濤;確定組合預測權(quán)系數(shù)最優(yōu)近似解的方法研究[J];系統(tǒng)工程理論與實踐;2000年03期

10 后銳;張畢西;;基于MLP神經(jīng)網(wǎng)絡的區(qū)域物流需求預測方法及其應用[J];系統(tǒng)工程理論與實踐;2005年12期

本文編號：2065155