Associate Professor Rensselaer Polytechnic Institute
Disclosure: Disclosure information not submitted.
Methods are needed to efficiently fulfill localized and elastic requests using a supplier base that has some choice as to which requests they accept. In hopes to achieve both high platform performance and supplier autonomy, we offer a Stackelberg-game-like hierarchical approach, with the platform as the leader, sifting through requests and sending a personalized set of options to each supplier. In response, the followers (suppliers) will either select a request to fulfill or choose to not fulfill any of the requests. A single time period version of this problem is captured as a mixed-integer linear bilevel optimization problem. The model determines the optimal request menu to send to each supplier to maximize revenue from fulfilled requests as well as to minimize the number of requests that are either rejected by all suppliers or are accepted by multiple suppliers. To address the uncertainty in supplier utility, we consider stochastic optimization approaches to our bilevel problem. A robust version of our model can be reformulated as a single-level MILP with complementarity constraints. We also explore classical stochastic approaches to analyze expected outcomes.