Finding Lower Bounds for Nondeterministic State Complexity Is Hard
We investigate the following lower bound methods for regular languages: The fooling set technique, the extended fooling set technique, and the biclique edge cover technique. It is shown that the maximal attainable lower bound for each of the above mentioned techniques can be algorithmically deduced from a canonical finite graph, the so called dependency graph of a regular language. This graph is very helpful when comparing the techniques with each other and with nondeterministic state complexity. In most cases it is shown that for any two techniques the gap between the best bounds can be arbitrarily large. Moreover, we show that deciding whether a certain lower bound w.r.t. one of the investigated techniques can be achieved is in most cases computationally hard, i.e., PSPACE-complete and hence is as hard as minimizing nondeterministic finite automata.
Unable to display preview. Download preview PDF.
- 1.Adorna, H.N.: Some descriptional complexity problems in finite automata theory. In: Saldaña, R.P., Chua, C. (eds.) Proceedings of the 5th Philippine Computing Science Congress, Cebu City, Philippines, March 2005, pp. 27–32. Computing Society of the Philippines (2005)Google Scholar
- 2.Bezrukov, S., Fronček, D., Rosenberg, S.J., Kovář, P.: On biclique coverings (preprint, 2005)Google Scholar
- 4.Brzozowski, J.A.: Mathematical theory of automata. In: Canonical Regular Expressions and Minimal State Graphs for Definite Events. MRI Symposia Series, vol. 12, pp. 529–561. Polytechnic Press, NY (1962)Google Scholar
- 7.Hopcroft, J.E., Ullman, J.D.: Formal Languages and Their Relation to Automata. Addison-Wesley, Reading (1968)Google Scholar
- 10.Hromkovič, J., Karhumäki, J., Klauck, H., Schnitger, G., Seibert, S.: Measures of nondeterminism in finite automata. Report TR00-076, Electronic Colloquium on Computational Complexity (ECCC) (2000)Google Scholar