The paper presents a multi-stage stochastic programming formulation for the planning of clinical trials in the pharmaceutical research and development (R&D) pipeline. Scenarios are used to account for the endogenous uncertainty in clinical trial outcomes. Given a portfolio of potential drugs and limited resources, the model determines the trials to be performed in each planning period and scenario. To reduce the size of the formulation we employ a reduced set of scenarios without compromising the quality of uncertainty representation. Furthermore, we present a number of ideas that allow us to reduce the number of non-anticipativity constraints necessary to model indistinguishable scenarios. The proposed approach is the first stochastic programming formulation to address this problem.
All Science Journal Classification (ASJC) codes
- Chemical Engineering(all)
- Computer Science Applications
- Mixed-integer programming
- Optimization under uncertainty
- Pharmaceutical research and development
- Stochastic programming