In the semantic web area, one of the most challenging issues involves computing the semantic similarity between specific terms. The problem here is the lack of precise dictionaries of the domain, e.g. biomedical, economic or any other. In this article, we recommend a new technique using various existing techniques of semantic similarity to achieve unique effects in the biomedical domain. In particular, we have built an evolutionary algorithm that makes use of the data provided through unique semantic similarity metrics. Our results were validated using several biomedical data sets.