Optimal Timing of Influenza Vaccine Production Under Yield-Improving and Knowledge Acquisition Strategies
42 Pages Posted: 21 Jun 2019
Date Written: June 16, 2019
In this paper, we consider a decision faced each year by an influenza vaccine manufacturer: when to optimally start large-scale production of the influenza vaccine after receiving viral candidate strains from the FDA. Prior to launching production, the manufacturer can take the time to stochastically increase the uncertain yield of the influenza vaccine by learning through experimentation how to improve virus growth conditions, thus forsaking immediate profits but raising future ones. After starting large-scale production, and receiving revenues from the sale of the vaccine, the manufacturer can continue to stochastically improve the vaccine yield by acquiring knowledge from real-time production data. We formulate and solve a multi-period, sequential-decision model to determine the optimal vaccine production timing while incorporating the dynamic evolution of vaccine yield uncertainty under those two knowledge acquisition strategies available to the vaccine manufacturer. We establish the structure of the optimal stopping policy for the timing of influenza vaccine production even though, due to the multifaceted impact of yield uncertainty on vaccine production economics, the resulting objective function is generally neither concave nor convex nor monotonic. We show that the structure of the optimal stopping policy depends in a fundamental way on the relative stochastic rates of the two knowledge acquisition strategies. We also analyze the implications of those strategies for bringing the influenza vaccine to the market earlier in the season. Using our results, we quantify the benefits from knowledge acquisition for increasing profits for the vaccine manufacturer and bringing the influenza vaccine to the market sooner. We discuss the resulting managerial implications.
Keywords: dynamic programming, optimal stopping, influenza vaccine, yield uncertainty
Suggested Citation: Suggested Citation