Genetic algorithms have proven their ability in detecting optimal or closed-to-optimal solutions to hard combinational problems. However, determining which crossover, mutation or selector operator is best for a specific problem can be cumbersome. The possibilities for enhancing genetic operators are discussed herein, starting with an analysis of their run-time performance. The contribution of this paper consist of analyzing the performance gain from the dynamic adjustment of the genetic operators, with respect to overall performance, as applied for the task of quantum circuit synthesis. We provide experimental results demonstrating the effectiveness of the approach by comparing our results against a traditional GA, using statistical significance measurements.
展开▼
