The focus of the work of this thesis is to develop demand uncertainty models for retailers making optimal pricing and inventory decisions on substitutable and perishable products. In particular we study three applications of demand uncertainty models: (i) a stochastic programming approach for two substitutable and perishable products over a two period planning horizon; (ii) a stochastic programming approach for multiple substitutable and perishable products over multiple periods; (iii) a robust optimization approach for two substitutable and perishable products over a single period. The three models support decision makers in retailing to incorporate the future demand uncertainty and substitution between similar products into their pricing and inventory decisions. In the context of a stochastic programming approach for two substitutable and perishable products problem over two periods, a stochastic dynamic programming model has been proposed in which the retailer aims at maximizing the total profit. The property of decision variables is analysed, an efficient search algorithm is developed to obtain the optimal results. Numerical results are reported using a case study based on a high-street fashion company. The sensitivity of the models' parameters is also analysed to address the great importance of data accuracy on decision variables and total profit. The benefits of considering pricing and inventory decisions simultaneously will be demonstrated and the total profit is observed to be significantly improved through the consideration of price substitution between substitutable products. In the context of a stochastic programming approach for multiple substitutable and perishable products problem over multiple periods, two stochastic dynamic programming models are proposed in which the decision maker can employ multiple markdowns on the prices. An efficient search algorithm has been developed by analysing the property of the decision variables. The benefits of making joint pricing and inventory decisions, considering substitutions between similar products; and dividing selling periods into more periods have been quantified. In the context of the robust optimization approach, we relax the assumption on the complete knowledge of the demand distribution from the stochastic dynamic programming model and develop a robust optimization model. The demand function is assumed to belong to an uncertainty set, and our objective is to find the optimal ordering quantity and price which maximize the worst-case profit. We extend a Newsvendor model in the face of uncertainty to consider the optimal pricing and inventory decisions of a retailer. Numerical tests are presented based on a case study of the retailing branch of a solar panel manufacturer. The trade-off between uncertainty level and total profits is illustrated, the sensitivity of parameters is also analysed.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:714513 |
| Date | January 2016 |
| Creators | Fang, Fei |
| Contributors | Nguyen, Tri-Dung |
| Publisher | University of Southampton |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | https://eprints.soton.ac.uk/399814/ |
Page generated in 0.002 seconds