联系我们

论文查询结果	返回搜索

论文编号：	9340
作者编号：	2120152411
上传时间：	2017/6/20 14:32:42
中文题目：	基于客户分类和销量预测的卷烟投放优化策略研究
英文题目：	Research on the Optimization Strategy of Cigarette Delivery based on Customer Classification and Sales Forecasting
指导老师：	李勇建
中文关键字：	卷烟投放优化策略；客户分类；产品需求分布；KKT条件
英文关键字：	Delivery Optimization Strategy; Customer Classification; Product Demand Distribution; KKT Method
中文摘要：	近年来，由于我国处于经济转型时期，烟草行业也处在适应经济转型调整的关键时期，烟草行业过去的高速发展也给烟草行业带来了巨大的挑战，市场发展对于行业的拉动力没有得到充分发挥，烟草企业以往的老旧经营管理模式难以适应现在全新的市场，行业发展面临瓶颈。随着市场的变化，卷烟的销售会受到市场需求的影响，为了适应市场经济体制的要求，烟草公司要根据市场需求制定合适的投放策略，从而在满足消费者需求的前提下，使得自身利润达到最大化。 本文使用云南烟草公司某州市的实际销售数据，采用合理的客户分类和预测方法。首先，现有的零售商是按照区域和经营业态进行分类的，但是同一区域同一经营业态的零售商所包含的消费群体可能会有不同的消费习惯。因此，按照这种方法进行分类不能完全展示出消费群体的行为特征，需要用更加科学的方法对零售商进行重新分类，因此采用合理的分类方法至关重要。本文运用改进的k-means方法对零售商进行分类。 其次，进行卷烟品牌的销量预测。在每一期制定策略时，需要知道当期消费者的需求，从而根据需求水平合理制定策略。因此，本文使用云南烟草公司某州市的历史销售数据，根据每一种卷烟品牌的历史销量得出品的需求分布，按照产品分布进行卷烟品牌的销量预测。 最后，根据新的卷烟零售商类别和产品的需求分布，制定出某州市卷烟品牌的投放策略。主要是基于该州市的历史数据，按照上级要求的指标制定出投放策略。首先是上级下达任务目标，明确本周期销量限制，然后根据任务目标将任务分给每个零售商，让各个零售商按照投放策略进行订购，使得烟草公司的收益达到最大。因此，该问题的目标函数就是烟草公司的收益最大化，同时会有很多的约束条件，比如销量限制、零售商资金限制等，在满足这些约束条件的前提下，求得最优的投放策略。此外，该优化模型属于带有约束条件的随机期望利润模型，可采用KKT条件对模型进行求解，然后将求解结果应用到实际的烟草公司销售案例中。
英文摘要：	In recent years, our country is in the period of economic transition, and the tobacco industry is in a critical period of adjustment of economic restructuring to adapt to the rapid development. The rapid development of the economic has also brought great challenges to the tobacco industry. Market development for the industry pulling power has not been fully put to good use. The old management mode of the tobacco enterprises is difficult to adapt to the new market, and the development is facing a bottleneck. With the change of the market, the sale of cigarettes will be affected by the market demand. In order to meet the requirement of the market economy system, the tobacco company needs to develop the appropriate marketing strategies according to the market demand, so as to meet the needs of consumers, to maximize it’s own profit. In this thesis, we use the actual sales data of the Yunnan tobacco company, through reasonable customer classification and forecasting methods to analyze the data. Firstly, existing retailers are classified according to the regional and the business formats. however, the same area of the same business retailers of the consumer groups may have different consumption habits. Therefore, classification according to this method can not fully show the behavior characteristics of consumer groups, retailers need to be reclassified with more scientific methods, so the reasonable classification method is very important. In this thesis, the k-means method is used to classify the retailers. Secondly, choosing the methods to forecast the cigarette brand sales. In each phase of the development strategy, you need to know the needs of the current consumer, in accordance with the level of demand to develop a reasonable strategy. Therefore, in this thesis, we get the distribution of the demand according to each kind of cigarette brand sales, and then, forecast the cigarette brand sales according to the distribution of the product. Finally, according to the demand of the new type of cigarette retail households and the distribution of products, the paper puts forward the strategy of cigarette brand in a state. Mainly based on the historical data of the city, according to the requirements of the higher authorities to develop a delivery strategy. The first is the task assigned by higher-up and determine the sales in this cycle limit, then assign tasks to each retailer according to the mission objectives. Each retailer orders in accordance with the launch strategy, making the tobacco companies reach the maximum revenue. Therefore, the objective function of the problem is to maximize the benefits of the tobacco companies, at the same time there will be a lot of constraints, such as sales restrictions, retail money restrictions, under the premise of satisfying these constraints, the optimal strategy is put forward. In addition, the optimization model is a stochastic expected profit model with constraints, which can be solved by KKT method, and then the results are applied to the actual tobacco companies sales case.
查看全文：	预览  下载(下载需要进行登录)