| dc.description.abstract | Short life produtcts retailing presents extra challenges when compared to retailing of non-perishable goods. Since these have a finite shelflife, all inventory must be sold until the product reaches its expiration date, when occour total or partial market value loss. In this context, retailers face the challenge of pricing their products, since high prices can generate losses due to obsolete inventory and low prices can reduce the revenue and consequently the company’s profit. Using tools that allow dynamic pricing of products based on the characteristics of their demand and their consumers can provide a competitive advantage to retail companies. The high volume of data needed and the complexity of this operation sometimes make its application unfeasible, but the advancement of technology and computational techniques can help retailers in this task. Among the techniques that can help this task is the machine learning algorithms. To solve this problem, a computational artifact was proposed combining the use of artificial neural networks for demand forecasting, a model for solving the newsvendor problem for sizing the inventory needed to meet the projected demand and a pricing model for markdown prices throughout the sales period, with the objective of maximizing the profit while minimizing losses due to excess inventory at the end of the sales season. The constructed artifact allowed the dynamic pricing practice from two propositions perspective, spaced discrete markdowns along the sales season for application in situations which the retailer deals with short-sighted customers and another of continuous markdowns following a costumer perceived value degeneration curve for situations which retailer deals with strategic behavior customers. In addition, the results indicate that its use can bring benefits to the sales campaigns planning, allowing more agility and precision in pricing. | en |
