Recent advancements in the biological sciences have created new problems for computational biology. One such problem is determining the sequence of DNA. One of the options for speeding up this process is to exploit the parallel nature of a 'sequencing by hybridisation' experiment to obtain the information from the physical structure of DNA. Various methods have been proposed to reconstruct DNA sequences from SBH data, including hybrid genetic algorithms. We describe a genetic algorithm which is more robust than previous methods, and gives results with a similar level of accuracy to alternative methods. The performance of the genetic algorithm is compared to alternative hybrid genetic algorithms, and improvements in the time the algorithm takes to return a result are recorded. We conclude by arguing that robust genetic methods may offer techniques to further improve on current methods.