Pseudo Boolean Optimization

Pseudo-Boolean optimization (PBO) focuses on finding optimal solutions for problems where the objective function and constraints are expressed as Boolean functions. Current research emphasizes developing efficient algorithms, including parallel local search methods, compact evolutionary algorithms like NSGA-II variants, and meta-solvers that leverage machine learning to select the best solver for a given problem and time constraint. These advancements aim to improve the scalability and performance of PBO solvers, impacting diverse fields such as machine learning (feature selection), hardware design (neural network optimization), and combinatorial optimization problems in general. The development of rigorous runtime analyses for various algorithms and problem classes further strengthens the theoretical foundation of the field.