Abstract
Incorporating biomass, a potential feedstock for chemical process industry, to our existing system will require significant investments on biomass conversion technologies. A recent model, a mixed integer nonlinear program (MINLP), addresses this investment planning problem. However, it has been observed that it becomes intractable quickly with the increase in the number of technologies. This paper investigates specialized solution approaches by exploiting the MINLP structure. Replacing bilinear terms with linearly segmented tight relaxations and nonlinear terms with linearly segmented under and upper estimators, solving the resulting MIP, and initializing DICOPT at the MIP solution solves the original MINLP efficiently.
