An adaptive parallel algorithm for finite language decomposition

Abstract The computationally hard problem of finite language decomposition is investigated. A L decomposable if there are two languages 1 and 2 such that = . Otherwise, prime. main contribution the paper an adaptive parallel algorithm for finding all decompositions based on exhaustive search incorporates several original methods pruning space. Moreover, since it changes its behavior runtime acquired data related to performance. Comprehensive computational experiments more than 4000 benchmark generated over alphabets various sizes have been carried out. showed by using power computing containing 200000 words can be found. Decompositions size not reported in literature so far.