In this talk, I present two research problems in radiation therapy, which we tackle using robust optimization and inverse optimization methods, respectively. The first problem involves designing optimized treatments that account for changes in the (uncertain) breathing motion pattern of a patient, which is a concern in the treatment of lung cancer. The second problem involves inferring objective function weights from historical treatments so as to guide the design of future treatments in a more data-driven (and less trial-and-error) manner. Both methodological and applied contributions will be presented.
Group for Research in Decision Analysis