One-Reversal Counter Machines and Multihead Automata: Revisited

Among the many models of language acceptors that have been studied in the literature are multihead finite automata (finite automata with multiple one-way input heads) and 1-reversal counter machines (finite automata with multiple counters, where each counter can only "reverse" once, i.e., once a counter decrements, it can no longer increment). The devices can be deterministic or nondeterministic and can be augmented with a pushdown stack. We investigate the relative computational power of these machines. Our results (where C1 and C2 are classes of machines) are of the following types: 1. Machines in C1 and C2 are incomparable. 2. Machines in C1 are strictly weaker than machines in C2. In obtaining results of these types, we use counting and "cut-and-paste" arguments as well as an interesting technique that shows that if a language were accepted by a device in a given class, then all recursively enumerable languages would be decidable.