• Fast, reliable implementations of many state-of-the-art multiobjective evolutionary algorithms • Extensible with custom algorithms, problems and operators • Supports master-slave, island-model, and hybrid parallelization • Tools for building and statistically testing new optimization algorithms • Professional support for businesses • Permissive open source license • Fully documented and tested source code

•Structured object oriented architecture •Smart pointers for automatic memory allocation management •Abstract factory, for runtime allocation of objects of any types •Parameters and algorithms dynamically configurable by files •Sophisticated logging mechanism with output in XML •Checkpoint mechanism for evolution recovery and results analysis •Population composed of multiple demes •History of best-of-run individuals for the whole population and for each demes •Population seeding from file •Master-slave model for parallel fitness evaluation