本文關鍵詞： 蝙蝠算法 BP神經(jīng)網(wǎng)絡 汽車配件 需求預測　出處：《西南交通大學》2017年碩士論文　論文類型：學位論文

【摘要】：隨著汽車產(chǎn)業(yè)的迅速發(fā)展,企業(yè)面臨著更為復雜的環(huán)境和更為強勁的競爭對手。因此,汽車企業(yè)不僅僅需要提升自身的制造技術,更要提高售后服務質(zhì)量。想要更穩(wěn)定,更優(yōu)質(zhì)的售后服務,汽車配件庫存量相對就要增多,但庫存量的增多,帶來的就是成本增加,庫存過剩的風險加大。因此,有效的、準確的汽車配件需求預測不僅能有效降低庫存的成本,還能提高汽車售后服務的質(zhì)量,使企業(yè)獲得更大的利益。首先,本文致力于設計并實現(xiàn)一個合理的汽車配件需求預測系統(tǒng)。以制造廠為核心,通過對汽車配件供應鏈和主要業(yè)務流程的研究,尋找影響汽車配件需求的關鍵因素。根據(jù)實際的調(diào)研需求,對比國內(nèi)外汽車配件預測方案,選擇由BA(蝙蝠算法)優(yōu)化BP神經(jīng)網(wǎng)絡的方式來建立預測模型,并通過模型的訓練和測試結(jié)果來驗證該模型的有效性。其次,預測的基礎在于數(shù)據(jù),傳統(tǒng)的配件需求預測往往只針對本地的數(shù)據(jù)進行預測,而忽視整個配件供應鏈上各節(jié)點數(shù)據(jù)之間的相互影響,這樣往往會導致需求預測不準確。本系統(tǒng)利用JAX-WS框架來開發(fā)WebService,結(jié)合RSA加密算法來實現(xiàn)系統(tǒng)與企業(yè)之間安全的數(shù)據(jù)交換,并用整合過來的數(shù)據(jù)進行模型訓練,提高模型預測的準確度。最后,根據(jù)系統(tǒng)需求實現(xiàn)系統(tǒng)管理模塊、基礎數(shù)據(jù)查詢模塊、數(shù)據(jù)交換模塊、配件預測查詢模塊和模型控制模塊,并完成對系統(tǒng)的功能測試和性能測試,驗證系統(tǒng)的準確性和穩(wěn)定性。本文將數(shù)據(jù)交換技術、預測技術和實際需求相結(jié)合,提高了預測的準確度和系統(tǒng)的可行性,并向用戶提供簡單、明了的可視化界面及多維度圖表的展示,讓用戶能更直觀的了解數(shù)據(jù),為其后期的需求決策提供幫助。
[Abstract]:With the rapid development of automobile industry, enterprises are faced with more complex environment and more powerful competitors. Therefore, automobile enterprises not only need to improve their manufacturing technology. More to improve the quality of after-sales service. Want more stable, more high-quality after-sales service, automotive spare parts inventory will be relatively increased, but the increase in inventory, it is cost increase. Therefore, effective and accurate demand prediction of auto parts can not only effectively reduce the cost of inventory, but also improve the quality of automobile after-sales service, so that enterprises can obtain greater benefits. This paper is devoted to design and implement a reasonable demand forecasting system for auto parts, taking the manufacturer as the core, through the research of the automobile parts supply chain and the main business process. In order to find the key factors that affect the demand of automobile parts, according to the actual research demand, comparing the domestic and foreign prediction schemes of auto parts, we choose the method of optimizing BP neural network by BA( bat algorithm) to build the prediction model. And through the model training and testing results to verify the effectiveness of the model. Secondly, the basis of the prediction is data, the traditional forecasting of accessories demand often only for the local data forecast. But ignore the interaction between the data of each node in the whole spare parts supply chain, which often lead to inaccurate demand prediction. This system uses JAX-WS framework to develop WebService. RSA encryption algorithm is used to realize the secure data exchange between the system and the enterprise, and the integrated data is used to train the model to improve the accuracy of model prediction. The system management module, basic data query module, data exchange module, accessory prediction query module and model control module are implemented according to the system requirements, and the function test and performance test of the system are completed. This paper combines data exchange technology, prediction technology and actual demand to improve the accuracy and feasibility of the system, and provide users with simple. The clear visual interface and the display of multi-dimensional charts enable users to understand the data more intuitively and provide help for their later demand decisions.
【學位授予單位】：西南交通大學
【學位級別】：碩士
【學位授予年份】：2017
【分類號】：TP311.52

【參考文獻】

相關期刊論文 前10條

1 周磊;張杰;;基于BA-BP開關磁阻電機電流分段建模研究[J];中國農(nóng)機化學報;2016年08期

2 王子毅;張春海;;基于ECharts的數(shù)據(jù)可視化分析組件設計實現(xiàn)[J];微型機與應用;2016年14期

3 王h熀，

本文編號：1470607