Robust encodings in genetic algorithms: a survey of encoding issues

Problems of encoding brittleness have been observed in the genetic algorithm (GA) literature, where slightly different problems require completely different genetic encodings for good solutions to be found. As research continues into GA encoding schemes the idea of encoding robustness becomes more important. A robust encoding is one which will be effective for a wide range of problem instances that it was designed for. A robust encoding will also be amenable to modification or extension to solve different problem types. This paper is a survey of some of the practical and theoretical considerations vital to the construction of a more robust encoding which will allow the GA to solve a broader range of problem types