The shortest common supersequence problem is a classical problem with many applications in different fields such as planning, Artificial Intelligence and especially in Bioinformatics. Due to its NP-hardness, we can not expect to efficiently solve this problem using conventional exact techniques. This paper presents a heuristic to tackle this problem based on the use at different levels of a probabilistic variant of a classical heuristic known as Beam Search. The proposed algorithm is empirically analysed and compared to current approaches in the literature. Experiments show that it provides better quality solutions in a reasonable time for medium and large instances of the problem. For very large instances, our heuristic also provides better solutions, but required execution times may increase considerably.
Gallardo, J. E. (2012). A Multilevel Probabilistic Beam Search Algorithm for the Shortest Common Supersequence Problem. PLoS ONE, 7(12). https://doi.org/10.1371/journal.pone.0052427