Advancements In Mixed Interger Programming For Project Portfolio Selection- A Case Study At Bosch

Abstract

A mixed-integer polynomial program with difficult non-convex cross-product terms is how the Project Portfolio Selection Problem (PPSP) is typically expressed. The existing approaches deal with this complexity by using a variety of linearization strategies, which are then followed by a branch-and-bound plan for calculations to get an accurate optimum solution. With a view to achieving a global optimum, this work presents an effective representation for PPSP that uses fewer continuous variables than previous approaches. Numerical tests are carried out to demonstrate the efficacy and efficiency of the suggested approach. Moreover, the approach is easily combined with a universal binary cut scheme, which makes it easier to find every possible answer. Using this method gives decision-makers the ability to think through a variety of choices for better decision-making.