Non-Stationary Demand Inventory Control Of Perishable Products

Perishable products is a type of commodities which have salient features such as long lead time, short sales period, unsold in sales end with a very low residual value and even to deal with costs, large uncertainty of market demand. Retailers often generate marketing strategies corresponding to the features of perishable products that continuous depreciation over time, leading to consumer demand for such commodities is divided into two stages, known as non-stationary demand. In the initial market, the demand is random. In the sales end, if retailers maintain the original price, commodity sales may reduce or even unmarketable. If retailers can not complete commodities sales on time, they will lose the cost of surplus commodities, or to bear high cost of long-term inventory expenses and occupation of capital. Therefore, at this stage retailers will use promotional stratagems such as discount. The study found that in this stage, market demand is related to commodity prices and inventory level. For retailers, how many commodities should be ordered in early sales, how to set a reasonable initial price, and how to set discount price dynamically in sales end, making their losses minimize or profits maximize while stimulating demand? These questions are practical problems which this thesis attempts to address.First, divide perishable products into three categories (perishable products, seasonal products and festive products), and summarize their own characteristics plus common characteristics of three categories of commodities, demand characteristics and the relevant technology and its application in non-stationary demand inventory control. Then, research into the two-stage non-stationary demand inventory control problem of seasonal products. The first stage, research the customer perceived value, customer arrival rate and customer total amount and other factors, set the potential demand to obey binomial distribution. To reduce inventory risk as the starting point, search for the optimal initial order quantity and initial pricing. The second stage, put forward the concept of Comprehensive Loss Rate, take account of comprehensive loss of quantity and time-value of money, set market demand dependent on commodity price and inventory level. The total planned time horizon is divided into several equal intervals. To minimize the loss of profits as the objective function, a two-stage dynamic ordering and pricing model of seasonal products based on time-varying comprehensive loss rate is established, solved by Particle Swarm Optimization (PSO), an applied numerical example is also given. Analysis about inventory changes and pricing strategies of two cases that the commodities can be or cannot be returned to the manufacturer are presented. Finally, summarize main contributions and innovations of this thesis and introduce the direction of future research.