發(fā)布時(shí)間：2018-05-17 19:50

  本文選題：產(chǎn)品組選擇 + 在線網(wǎng)評(píng)��； 參考：《西南財(cái)經(jīng)大學(xué)》2014年碩士論文

【摘要】：產(chǎn)品組選擇問題是存在于零售行業(yè)、航空業(yè)、家用電子行業(yè)等行業(yè)內(nèi)的基礎(chǔ)性問題,也是主要的運(yùn)作問題。產(chǎn)品組選擇即：商家在資金、倉儲(chǔ)、店面等資源有限的情況下從數(shù)量眾多的產(chǎn)品中選擇一組產(chǎn)品來進(jìn)行銷售的過程。目的旨在通過對(duì)這些限制條件的權(quán)衡達(dá)到產(chǎn)品銷量的提升和利潤的最大化。零售商希望在合適的時(shí)間,合適的地點(diǎn),提供合適的商品組合來滿足消費(fèi)者的需求。首先,做到這點(diǎn)不容易,其次,真的做到了這一點(diǎn)也并不能保證商家的產(chǎn)品組選擇一定正確,因?yàn)楫a(chǎn)品組選擇是由每一個(gè)銷售時(shí)點(diǎn)以及每一個(gè)零售店具體的銷售情況決定的,所以零售商應(yīng)該具體問題具體分析,不斷調(diào)整產(chǎn)品策略來適應(yīng)季節(jié)的變遷、新產(chǎn)品的引入、消費(fèi)者喜好的改變等因素。所有因素里,無疑,經(jīng)濟(jì)環(huán)境的影響是深遠(yuǎn)的。 在互聯(lián)網(wǎng)經(jīng)濟(jì)不斷發(fā)展的今天,越來越多的企業(yè)意識(shí)到網(wǎng)絡(luò)信息傳播的重要性,與網(wǎng)絡(luò)密切打交道的電商企業(yè)更是不例外。作為最近幾年蓬勃發(fā)展的一個(gè)新興行業(yè),電商企業(yè)已逐漸顯示出它的優(yōu)勢(shì)和生命力。網(wǎng)上零售商的經(jīng)營者享受著購物信息在消費(fèi)者之間充分流轉(zhuǎn)的同時(shí),必須時(shí)刻警惕網(wǎng)絡(luò)信息的負(fù)面影響,只有研究并關(guān)注網(wǎng)上消費(fèi)者的消費(fèi)行為,掌握他們的網(wǎng)評(píng)信息因素,才能化不足為長處,為自己的發(fā)展找到方向。 同傳統(tǒng)購物模式相比,網(wǎng)上消費(fèi)者的購買決策會(huì)更多受到網(wǎng)絡(luò)產(chǎn)品信息的影響,相對(duì)來說,他們會(huì)更加理性地比較各類網(wǎng)評(píng)信息,因?yàn)榫W(wǎng)絡(luò)購買受外界因素影響較小,購買者可以面對(duì)電腦屏幕更全面地瀏覽商品的網(wǎng)評(píng)信息。網(wǎng)評(píng)信息主要地包括網(wǎng)評(píng)得分,網(wǎng)評(píng)有用性,得分方差,網(wǎng)評(píng)文字部分,網(wǎng)評(píng)總量。也有很多學(xué)者就這些指標(biāo)對(duì)產(chǎn)品銷量的影響展開研究,然而,本文發(fā)現(xiàn),在關(guān)于網(wǎng)評(píng)維度的研究里,學(xué)者并沒有就某一項(xiàng)對(duì)于商家銷量的影響達(dá)成共識(shí),這使得網(wǎng)評(píng)研究具有廣闊發(fā)展空間的同時(shí),選取合適的網(wǎng)評(píng)指標(biāo)作為研究對(duì)象成為棘手之事。本文參照Sun(2010)和Nan Hu(2010)的研究觀點(diǎn),選取網(wǎng)評(píng)得分(Valence)和評(píng)分標(biāo)準(zhǔn)方差(Variance)這兩項(xiàng)網(wǎng)評(píng)維度作為研究零售商產(chǎn)品組選擇問題的影響因素。 而企業(yè)的資源是有限的,經(jīng)營者總是希望能夠通過正確地配置資源,達(dá)到收益的最大化。不同的產(chǎn)品各有特點(diǎn),如何利用好這些產(chǎn)品之間的組合選擇,特別是加入了無形的網(wǎng)評(píng)信息,這使得產(chǎn)品評(píng)估信息更加多元化,對(duì)諸多網(wǎng)上零售企業(yè)來說,做到合理的權(quán)衡是一件難事。消費(fèi)者對(duì)產(chǎn)品的決策是一個(gè)離散選擇的過程,因此離散選擇模型(Discrete Choice Model)在這類問題中得到廣泛應(yīng)用。本文選取最基礎(chǔ)也是運(yùn)用最廣泛的MNL (Multinomial Logit)模型來模擬顧客的選擇過程。長久以來,在零售商產(chǎn)品組合優(yōu)化問題中,解空間是隨著可選產(chǎn)品集合數(shù)量成指數(shù)增長的,這是一個(gè)NP-Hard司題,一般很難精確地求出最優(yōu)解,因此尋求一個(gè)有效的近似解算法是十分必要的。實(shí)踐證明,遺傳算法正是啟發(fā)式算法里應(yīng)用比較廣泛、求解比較得力的一種算法,但針對(duì)不同的優(yōu)化問題,也會(huì)存在嚴(yán)重的局限性,尤其表現(xiàn)在產(chǎn)品組合優(yōu)化的求解問題中。因此本文采用的是改進(jìn)的基于貪心策略的遺傳算法,以期望能夠克服單純遺傳算法的某些弱點(diǎn),來更快達(dá)到最優(yōu)解。 追溯相關(guān)產(chǎn)品組選擇問題的研究,本文發(fā)現(xiàn),研究主要集中在庫存與產(chǎn)品組選擇的關(guān)系,缺貨替代與消費(fèi)者需求之間的關(guān)系,產(chǎn)品組選擇的最優(yōu)算法或者近似最優(yōu)算法等,沒有學(xué)者將消費(fèi)者對(duì)于產(chǎn)品的網(wǎng)評(píng)信息納入到產(chǎn)品組選擇問題中來。因此,在網(wǎng)絡(luò)經(jīng)濟(jì)時(shí)代,筆者認(rèn)為將二者結(jié)合起來,做出嘗試性的探討是有意義的。本文的創(chuàng)新之處主要體現(xiàn)在：1.在離散選擇模型的基礎(chǔ)上加入了除傳統(tǒng)的價(jià)格影響因素外的網(wǎng)評(píng)得分與標(biāo)準(zhǔn)方差的新元素,以適用于網(wǎng)絡(luò)環(huán)境的新背景,提出了具體的消費(fèi)者需求評(píng)估模型,即改進(jìn)版的MNL模型,拓寬了MNL模型的理論應(yīng)用范圍；2.創(chuàng)新性地將產(chǎn)品分類計(jì)劃的研究問題考慮到時(shí)下熱門的電子商務(wù)環(huán)境中來,試圖發(fā)現(xiàn)消費(fèi)者網(wǎng)評(píng)信息對(duì)于電商產(chǎn)品組選擇策略的影響,希望能夠做出一點(diǎn)有益的幫助。在導(dǎo)師的指導(dǎo)下,本人通過使用改進(jìn)的MNL離散選擇模型,將顧客效用表達(dá)式內(nèi)涵地包括了產(chǎn)品價(jià)格、網(wǎng)評(píng)得分與評(píng)分標(biāo)準(zhǔn)方差,并使用基于貪心策略的遺產(chǎn)算法將零售商產(chǎn)品組選擇問題轉(zhuǎn)化為生物遺傳語言,進(jìn)行MATLAB編程,通過模型的構(gòu)建,分析了不同參數(shù)設(shè)置下的結(jié)果變化情況,得出了如下結(jié)論： 1.產(chǎn)品的網(wǎng)評(píng)得分對(duì)于電商的產(chǎn)品組選擇有著積極,正向的影響,顧客越在意產(chǎn)品的網(wǎng)評(píng),也就越愿意選擇網(wǎng)評(píng)得分高的產(chǎn)品,這樣就會(huì)有更多的顧客將資金用在幾種評(píng)分高的產(chǎn)品,這就導(dǎo)致了雖然選擇的產(chǎn)品少了,卻會(huì)有更大的利潤。建議零售商在顧客逐漸重視網(wǎng)評(píng)得分的情況下,應(yīng)該逐漸減少產(chǎn)品的種類,以便于顧客購買產(chǎn)品。 2.網(wǎng)評(píng)得分標(biāo)準(zhǔn)方差與產(chǎn)品組選擇之間呈現(xiàn)出負(fù)向的關(guān)系,也就是說,隨著消費(fèi)者對(duì)方差影響敏感系數(shù)的增大,商家的利潤減小,產(chǎn)品種類數(shù)增大。這說明如果顧客逐漸看重商品網(wǎng)評(píng)評(píng)分的集中程度,那么消費(fèi)者選擇商品的概率就減小。商家為了得到更多的利潤,就不得不增加產(chǎn)品的種類。 3.關(guān)于算法實(shí)現(xiàn),在本文算例中證明,遺傳算法中種群大小的改變會(huì)或多或少影響最終商家選擇結(jié)果的變化。在其逐漸變大的過程中,遺傳進(jìn)化代數(shù)總體變少,但每一次迭代的速度變慢了,收斂時(shí)間變長了,總體能在在較少時(shí)間內(nèi)找到問題最優(yōu)解。 本文在網(wǎng)絡(luò)時(shí)代消費(fèi)者日益受到網(wǎng)評(píng)信息影響的背景下,構(gòu)建了在線評(píng)論影響商家產(chǎn)品選擇的模型,研究了在線評(píng)論對(duì)網(wǎng)絡(luò)零售商產(chǎn)品決策的影響,建議了電商企業(yè)重視相關(guān)網(wǎng)評(píng)信息的收集與評(píng)估,豐富了PAP問題的研究范圍,具有一定的創(chuàng)新性。但是由于時(shí)間和能力的制約,論文無法全面分析在線評(píng)論對(duì)商家決策的影響,只能從局部的幾點(diǎn)出發(fā),具有一定的片面性。 盡管論文關(guān)于在線評(píng)論對(duì)電商企業(yè)產(chǎn)品組選擇的影響的研究不夠深入透徹,但畢竟是一次有益的嘗試,希望能為以后的深入研究提供有效的參考,并具備一定的理論價(jià)值和實(shí)踐意義。
[Abstract]:The choice of product group is the basic problem that exists in the retail industry, the aviation industry, the home electronics industry and so on. The product group chooses to choose a group of products from a large number of products in the case of limited resources such as funds, warehouses and stores. The trade-off between these constraints can achieve the promotion of product sales and the maximization of profit. The retailer wants to meet the needs of the consumer at the right time, the right place, and the suitable product mix. First, it is not easy to do this. Secondly, it does not guarantee the choice of the product group of the merchant to be correct. Because the selection of the product group is determined by each time point of sales and the specific sales situation of each retail store, the retailer should analyze the specific problems and adjust the product strategy to adapt to the changes of the season, the introduction of new products, the change of consumer preferences, and so on. All factors, undoubtedly, the impact of the economic environment is Profound.
With the continuous development of the Internet economy, more and more enterprises are aware of the importance of the communication of network information. The business enterprise which is closely related to the network is no exception. As a booming industry in recent years, e-commerce enterprises have gradually shown its advantages and vitality. The operators of online retailers enjoy it. While shopping information is fully transferred among consumers, we must always be vigilant on the negative effects of network information. Only studying and paying attention to consumer behavior on the Internet and mastering the information factors of their online reviews can make the lack of the strengths and find the direction for their own development.
Compared with the traditional shopping model, the purchase decision of online consumers will be more affected by the information of the network products. Relatively speaking, they will compare all kinds of online reviews more rationally, because the network purchase is less influenced by the external factors and the buyers can browse the online reviews more comprehensively on the computer screen. The key points include the score of net reviews, the usefulness of the net reviews, the variance of the score, the text of the net reviews, the total amount of the net reviews. There are also many scholars on the impact of these indicators on the sales of products. However, this paper finds that in the study of the dimensions of the net reviews, the scholars have not reached consensus on the impact of a particular item on the sales of the business, which makes the online review study At the same time, it is a thorny thing to choose the suitable index of net reviews as the research object. This paper, referring to the research views of Sun (2010) and Nan Hu (2010), selects the two dimensions of net evaluation score (Valence) and score standard variance (Variance) as the influencing factors of the study on the selection of the retailer's product group.
And the resources of the enterprise are limited, and the operators always hope to maximize the income by correctly configuring the resources. The different products have their own characteristics, how to make use of the combination selection between these products, especially the invisible net evaluation information, which makes the product evaluation information more diversified and to many online retail enterprises. It is difficult to make a reasonable tradeoff. The decision of the consumer to the product is a discrete selection process, so the discrete selection model (Discrete Choice Model) is widely used in this kind of problem. This paper selects the most basic and most widely used MNL (Multinomial Logit) model to simulate the process of customer selection. In the problem of retailer portfolio optimization, the solution space is increasing exponentially with the number of optional product sets. This is a NP-Hard division problem. It is generally difficult to find the optimal solution accurately. Therefore, it is necessary to seek an effective approximate solution algorithm. Practice has proved that the genetic algorithm is widely used in heuristic algorithms. In order to solve the problem of different optimization problems, there is also a serious limitation, especially in the solution of product combination optimization. Therefore, this paper uses an improved genetic algorithm based on greedy strategy to overcome some weaknesses of the simple genetic algorithm to achieve the optimal solution.
It is found that the study mainly focuses on the relationship between inventory and product group selection, the relationship between the stock replacement and the consumer demand, the optimal algorithm or the approximate optimal algorithm for the selection of the product group, and no scholars incorporate the consumer's information on the product to the product group selection problem. Therefore, in the era of network economy, I think it is meaningful to combine the two and make an attempt to make an attempt. The innovation of this article is mainly embodied in the following: 1., on the basis of the discrete choice model, the new elements of the net assessment score and the standard variance are added to the network environment, which is suitable for the new back of the network environment. In view, a specific consumer demand assessment model, that is, an improved MNL model, has been developed to broaden the scope of the theoretical application of the MNL model. 2. the research issues of the product classification plan are innovatively taken into consideration in the popular e-business environment, and the effect of consumer online reviews on the selection strategy of ecommerce product group is sought. Under the guidance of the tutor, by using the improved MNL discrete selection model, the customer utility expression includes the product price, the net assessment score and the standard variance, and the legacy algorithm based on the greedy strategy transforms the retailer product group selection problem into the biological genetic language. MATLAB programming is carried out, and the results of different parameters are analyzed through the construction of the model. The following conclusions are drawn:
The net evaluation score of 1. products has a positive and positive impact on the selection of the product group. The more customers care about the product online reviews, the more they are willing to choose the products with high score in the online review. In this way, more customers will be able to use the funds in several high grade products, which leads to a greater profit, although the selection of products is less. It is suggested that retailers should gradually reduce the types of products in order to facilitate customers to buy products when customers gradually pay attention to scoring.
2. there is a negative relationship between the standard variance of score and the selection of product group, that is to say, with the increase of the sensitivity coefficient of the consumers' difference, the profit of the merchant is reduced and the number of products increases. This shows that the probability of consumer choice of goods is reduced if the customer gradually emphasizes the concentration range of the net evaluation score of the commodity. In order to get more profits, businesses will have to increase the variety of products.
3. on the implementation of the algorithm, it is proved in the example of this paper that the change of the population size in the genetic algorithm will more or less affect the change of the final merchant selection result. In the process of increasing the population, the genetic evolution algebra is less generally, but the speed of each iteration is slower, the convergence is longer, and the general can find the question in less time. The optimal solution of the problem.
In this paper, under the background that consumers are increasingly affected by the information of online reviews in the network era, the model of online reviews affecting product selection is constructed, and the impact of online reviews on product decision making of network retailers is studied. It is suggested that e-commerce enterprises attach importance to the collection and evaluation of related network evaluation information, which enriches the scope of research on PAP problems and has a certain extent. However, due to the constraints of time and ability, the paper can not fully analyze the impact of online reviews on business decision making, which can only start from a few points and have a certain one-sided nature.
Although the research on the impact of online reviews on the selection of e-commerce enterprise product group is not thorough enough, it is a useful attempt after all, hoping to provide an effective reference for further research, and have certain theoretical and practical significance.
【學(xué)位授予單位】：西南財(cái)經(jīng)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：F224;F724.6

【參考文獻(xiàn)】

相關(guān)期刊論文 前8條

1 苗蘊(yùn)慧;唐加福;張鐵軍;;基于參考依賴的改進(jìn)MNL模型預(yù)測(cè)電信套餐選擇行為[J];系統(tǒng)工程;2013年06期

2 戴曉明,許超,龔向陽,邵惠鶴;并行遺傳算法收斂性分析及優(yōu)化運(yùn)算[J];計(jì)算機(jī)工程;2002年06期

3 張曉繢，戴冠中，，徐乃平;一種新的優(yōu)化搜索算法──遺傳算法[J];控制理論與應(yīng)用;1995年03期

4 趙宏立,龐小紅,吳智銘;基因塊編碼的并行遺傳算法及其在TSP中的應(yīng)用[J];上海交通大學(xué)學(xué)報(bào);2004年S1期

5 袁誠;離散選擇模型在產(chǎn)品差別研究中的應(yīng)用[J];統(tǒng)計(jì)研究;2003年01期

6 江雷;基于并行遺傳算法的彈性TSP研究[J];微電子學(xué)與計(jì)算機(jī);2005年08期

7 張群,曹麗,李純青;離散選擇模型的比較及其在零售業(yè)的應(yīng)用[J];西安工業(yè)學(xué)院學(xué)報(bào);2005年03期

8 白瑞;沈其君;張曉;唐立健;;嵌套logit模型原理及其在醫(yī)學(xué)離散選擇問題分析中的應(yīng)用[J];中國衛(wèi)生統(tǒng)計(jì);2006年02期

本文編號(hào)：1902632