Conference Agenda

The supply chain for mobile phone refurbishment is characterized by a highly stochastic environment. To maximize the long-term average profit of the refurbishment system, we derive the optimal control strategies for procurement, processing, and pricing. We propose a linear programming formulation for a Markov Decision Process that considers the uncertain arrival of multiple-raw materials. The effects of the variabilities of procurement, price and demand, of the processing rate, and of the customers’ price sensitivity on the performance are investigated. Our numerical experiments use industry data and lead to important managerial insights for our industry partner.

Amidst the current challenges of the food industry, including an ever-expanding product variety, optimizing planning processes within production becomes imperative. Due to a shortage of skilled workers, heterogeneous qualifications, and interdependencies between departments and employees, efficient workforce assignment is critical. This work studies a real-world worker assignment problem of a German organic food company. The main particularity is the consideration of a heterogeneous workforce regarding employees' qualifications and skill levels. Additionally, this work includes production-based time intervals as well as shift plans. An integer linear program (ILP) is introduced to assign the workforce optimally. The company's primary goal is to use its staff efficiently and avoid needing temporary workers due to insufficient staff or qualifications. Therefore, the objective function minimizes inefficient assignments and the use of temporary workers. The ILP is used to present an employee schedule to the personnel planners as real-world decision support. Additionally, the solutions point out the company's worker assignment potential, revealing staff utilization inefficiencies, qualification shortages, and imbalances.

Producers of aging cheese serve multiple demand streams for products with different maturation ages. The distinct taste of these age-differentiated cheeses prevents product substitution, posing unique challenges in production decisions. These involve balancing immediate revenue from selling younger cheeses against potentially higher future earnings from cheeses aged longer. Additionally, managers must deal with uncertainties, as raw milk costs and final product sales prices follow correlated stochastic processes.

Our Markov Decision Process formulation uses Ornstein-Uhlenbeck processes to capture the price dynamics. The action space includes purchasing decisions, i.e., the amount of raw milk transformed into young cheese, and production decisions, i.e., the volumes of different products placed in the sales market. Further, as product labels define age ranges such as “matured for 3-6 months”, issuance decisions determine the allocation of stock volumes from specific age classes to individual products.

We propose a novel Deep Reinforcement Learning algorithm that combines Average Policy Optimization with a rolling horizon lookahead heuristic. In numerical experiments, we investigate the effect of price process parameters on near-optimal policies. Our findings suggest that the value of using age ranges on product labels increases with the mean reversion rate of these processes.