Dynamic Resource Allocation with Proactive Upgrades : The Value and Design of Fulfillment Flexibility

Firms engage in product proliferation to take advantage of heterogeneous customers. Product upgrading, where a customer may be served with an upgrade to a higher quality product or service, has been an increasingly popular practice. Despite its popularity, there is limited understanding of how resources should be managed dynamically and how much value is added by providing upgrades under a general upgrade structure. This work attempts to close this gap by considering dynamic network revenue management with multiple vertically differentiated products under a general upgrades structure. We start with a simple network of three products and show that the optimal dynamic upgrading can be controlled by state-dependent protection levels, that are monotonic with limited sensitivity. For a general network with general upgrades, based on the structural property of the optimal policy, we focus on developing computationally efficient heuristics using fluid approximation. We show that the heuristic is asymptotically optimal with a large demand. We find that the upgrading flexibility could improve revenue substantially (more than 20\%) depending on the mismatch between supply and demand, the upgrading structure, the demand correlation, and the number of products. We also show that the first few levels of upgrading flexibility may not be effective, depending on demand correlation. Our findings provide useful guidance on the implementation and design of upgrading flexibility for practitioners under general upgrading structures