Multi-Objective Evolutionary Optimization

The field studies how to find sets of solutions that capture trade-offs among competing goals, rather than a single best answer. Common approaches include NSGA-II and NSGA-III, SPEA2, MOEA/D, and differential evolution, with diversity preservation and constraint-handling methods to guide the search. Current work tackles large-scale, dynamic, uncertain, and expensive problems using techniques such as variable grouping, surrogates, adaptive operators, and learned guidance, across applications from engineering design to scheduling and machine learning. A central challenge is balancing convergence, diversity, feasibility, and evaluation cost; performance remains problem-dependent, while runtime theory and reliable algorithm selection are active concerns.