Polynomial-time parametric optimisation

Abstract: In biology, predicting RNA secondary structures plays a vital role in determining its physical and chemical properties. Although we have powerful energy models to predict them as well as parametric analysis to understand the models themselves, the large number of parameters involved makes exploring the parameter space and effective fine-tuning complicated at best. The literature describes an approach via so-called RNA polytopes and several attempts to compute them entirely, but computing explicitly the polytopes is both practically and theoretically intractable. In this thesis, we demonstrate how to further modify the dynamic programming algorithms used in RNA secondary structure prediction, and more generally how to use only supporting functions to gather some information about the polytopes without explicit construction. We provide the mathematical frameworks with proofs or sketch thereof whenever necessary, and carry out some numerical experiments to show that our proposed methods are practical even when the number of parameters is large. As it turns out, one of our methods provides a solution to another problem in computational geometry previously unsolved to our knowledge, and we hope this thesis will accommodate future studies in RNA, as well as inspire further researches on the potential uses of polytopes' supporting functions in computational geometry.