Modern data-driven inventory control methods of many stock-keeping-units

This 3-year PhD project is for candidates with an interest in multi-objective or multi-level stochastic non-linear optimization problems applied to the area of inventory control and product range/revenue management in support of firms that have to deal with many stock-keeping-units. Applications include retail chains that sell home furniture, electric and electronic appliances, do-it-yourself stores, spare parts stores of cars and bikes, and spare parts service organizations to both consumers and businesses.
The candidate will have the opportunity to work together alongside experts in both the mathematical challenges and the application areas. The ideal candidate will have a background in operational research, mathematics, statistics, or computer science and have good programming skills. The candidate will join the department of Mathematical Sciences alongside about 20 new PhD students. The aim of this project is to work on the mathematics of the problem – the specification of the mathematical models and algorithms to solve the models, not on delivering a bespoke decision support tool!

More about the project:
Inventory control deals traditionally with the questions of when and how much to order of a product so as to meet customer service requirements at minimal total logistics costs (including the transportation costs and the financial opportunity cost of investing in stocks). Firms in various industries typically deal with the inventory control of many thousands of individual Stock-Keeping-Units (SKUs). ABC classification is a popular approach by which these products are classified according to simple criteria such as demand value or demand volume, and such that the most sophisticated inventory control methods can be applied to the most important items of the range. However, this on itself does not specify the actual inventory control methods that should be deployed for each of the SKUs, nor accounts for constraints that link the decisions made for each SKU and that decisions taken at a higher level such as product range and promotional activities, also affect the outcome.

In a modern world where data-driven decision making becomes a reality, closer attention can be paid to exploiting these links between inventory control and other areas of decision making such as forecasting, product range management, and revenue optimization. Aspects that link the inventory control of a single item to that of other items and to other decisions include warehouse capacity constraints and decisions on store management and promotional runs.

The first aim of this project is to develop the mathematics behind a computer-based optimization decision tool for helping logistics and supply chain managers to make ordering decisions across the whole range of their SKUs that are cost minimizing, meet customer service level targets, and are supportive of the product range and revenue optimization strategies and tactics deployed by the firm. The next step is to further extend this to include also parts of the product range and promotional decisions into the decision method. This may lead to large-scale multi-objective or multi-level stochastic optimization models. The aim of this PhD project is to work on the mathematics of the problem, not on delivering a bespoke decision support tool.