UROP Project

Process Simulation and Global Optimization in MAiNGO

In previous work, we were able to show that Aspen, a commercial process simulator in chemical engineering, makes mistakes in thermodynamic calculations and declares incorrect results as error-free. These can be optimized to the thermodynamically correct solution with newly developed models in MAiNGO (Mc cormick-based Algorithm for Mixed-Integer Nonlinear Global Optimization, short, MAiNGO).

Task

You will implement and solve engineering related problems with our in-house solver, based on McCormick relaxations. Focus is on global optimization of constraint nonlinear problems in the field of process systems engineering and further development of our solver. Develop strategies for efficient computation of the Branch and Bound algorithm, its process engineering models with guaranteed thermodynamic stability (Semi Infinite Programming) and compare with available benchmarks. The implementation will be performed in C++ for single level and GAMS for multi-level optimization

Requirements

Student in the field: CES/ MechE/ Chemical Engineering/ Math/ Simulation Science. Interest in optimization is advantageous. Previous knowledge in C++/GAMS or process engineering is not mandatory but favorable. Please apply with transcript This Work is a preparation for students considering doing their PhD.