Details

Autor(en) / Beteiligte

• John, Stefan
• Kosiol, Jens
• Taentzer, Gabriele

Titel

Towards a configurable crossover operator for model-driven optimization

Ist Teil von

• Proceedings of the 25th International Conference on Model Driven Engineering Languages and Systems: Companion Proceedings, 2022, p.388-395

Ort / Verlag

New York, NY, USA: ACM

Erscheinungsjahr

2022

Link zum Volltext

Quelle

ACM Digital Library

Beschreibungen/Notizen

• In evolutionary algorithms, mutation and crossover are used to explore a search space for solutions. For the model-based approach to model-driven optimization, where models are used to represent solutions, no crossover operator has been introduced yet. However, theoretical and experimental evidence shows that evolutionary search can benefit from the use of crossover. We present a configurable crossover operator for models defined in the Eclipse Modeling Framework (EMF), discuss several variants of incorporating domain knowledge into this operator, and argue that it produces EMF models again. We also present a prototype implementation of our crossover operator and conduct an initial evaluation to investigate the effectiveness of evolutionary computations that use both mutation and crossover.