Abstract
We present new simulations for \(\text {ASpace}^{\mathrm {dm}}(s(n))\), the class of languages that can be accepted by alternating Turing machines starting with s(n) worktape cells delimited initially. Under weak constructibility assumptions, not excluding monotone functions below \(\log n\), we show: (i) \(\text {ASpace}^{\mathrm {dm}}(s(n))\subseteq \text {DTime}(n\!\cdot \!2^{O(s(n))})\). This extends, to sublogarithmic space, the classical simulation of alternating space by deterministic time. (ii) \(\text {ASpace}^{\mathrm {dm}}(s(n))\subseteq \text {NTimeSpace}(n\!\cdot \!2^{O(s(n))},2^{O(s(n))})\), a simulation with simultaneous bounds on time and space. This improves the known inclusion, stating that \(\text {ASpace}^{\mathrm {dm}}(s(n))\subseteq \text {NSpace}(2^{O(s(n))})\). (iii) \(\text {ASpace}^{\mathrm {dm}}(s(n))=\text {co-}\text {ASpace}^{\mathrm {dm}}(s(n)))\), i.e., the alternating space is closed under complement. This simulation does not depend on whether s(n) is above \(\log n\) nor on whether the original machine gets into infinite loops, which solves a long-standing open problem.
Supported by the Slovak grant contracts VEGA 1/0142/15 and APVV-15-0091.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
By \(X\text {Space}^{\mathrm {dm}}(s(n))\), for \(X\in \{\text {D},\text {N},\text {A}\}\), we denote the classes of languages accepted by deterministic, nondeterministic, and alternating Turing machines starting with a worktape consisting of \(\lfloor {s(n)}\rfloor \) blank cells delimited by endmarkers (here \(\lfloor {x}\rfloor \) denotes the largest integer satisfying \(i\le x\), for the given real value x), as opposed to the more common complexity classes \(X\text {Space}(s(n))\) where the worktape is initially empty and the machine must use its own computational power to make sure that it respects, along each computation path on each input of length n, the space bound of s(n). The notation “dm” derives from “Demon” Turing Machines [5].
- 2.
It is known that \(\text {ASpace}(o(\log \log n))\) contains only regular languages [15]. However, it is still possible to accept some nonregular languages, if \(\lfloor {s(n)}\rfloor \le o(\log \log n)\) worktape cells are delimited automatically at the very beginning. As an example [2], take \(\mathcal {L}= \{1^{n}:\,n\mod \lceil {\log \log \log n}\rceil =0\}\), contained in \(\text {DSpace}^{\mathrm {dm}}(\log \log \log \log n)\).
- 3.
For \(i\in \{0,n+1\}\), we take \(a_{i}\in \{\rhd ,\lhd \}\), two new symbols representing the respective endmarkers.
References
Allender, E., Mix Barrington, D., Hesse, W.: Uniform circuits for division: consequences and problems. In: Proceedings of IEEE Conference on Computer Complexity, pp. 150–159 (2001)
Bertoni, A., Mereghetti, C., Pighizzini, G.: On languages accepted with simultaneous complexity bounds and their ranking problem. In: Prívara, I., Ružička, P., Rovan, B. (eds.) MFCS 1994. LNCS, vol. 841, pp. 245–255. Springer, Heidelberg (1994)
Braunmühl, B., Gengler, R., Rettinger, R.: The alternation hierarchy for sublogarithmic space is infinite. Comput. Complex. 3, 207–230 (1993)
Chandra, A., Kozen, D., Stockmeyer, L.: Alternation. J. Assoc. Comput. Mach. 28, 114–133 (1981)
Chang, R., Hartmanis, J., Ranjan, D.: Space bounded computations: review and new separation results. Theoret. Comput. Sci. 80, 289–302 (1991)
Geffert, V.: A speed-up theorem without tape compression. Theoret. Comput. Sci. 118, 49–65 (1993)
Geffert, V.: A hierarchy that does not collapse: alternations in low level space. RAIRO Inform. Théor. Appl. 28, 465–512 (1994)
Geffert, V.: Bridging across the \(\log (n)\) space frontier. Inform. Comput. 142, 127–158 (1998)
Geffert, V., Pardubská, D.: Unary coded NP-complete languages in ASPACE\((\log \log n)\). Int. J. Found. Comput. Sci. 24, 1167–1182 (2013)
Hartmanis, J., Lewis II, P., Stearns, R.: Hierarchies of memory limited computations. In: IEEE Conference on Record on Switching Circuit Theory and Logical Design, pp. 179–190 (1965)
Hennie, F.: One-tape, off-line Turing machine computations. Inform. Control 8, 553–578 (1965)
Hong, J.: A tradeoff theorem for space and reversal. Theoret. Comput. Sci. 32, 221–224 (1984)
Hopcroft, J., Motwani, R., Ullman, J.: Introduction to Automata Theory, Languages, and Computation. Addison Wesley, Reading (2001)
Immerman, N.: Nondeterministic space is closed under complementation. SIAM J. Comput. 17, 935–938 (1988)
Iwama, K.: ASPACE\((o(\log \log n))\) is regular. SIAM J. Comput. 22, 136–146 (1993)
Liśkiewicz, M., Reischuk, R.: The sublogarithmic alternating space world. SIAM J. Comput. 25, 828–861 (1996)
Mereghetti, C.: Testing the descriptional power of small Turing machines on nonregular language acceptance. Int. J. Found. Comput. Sci. 19, 827–843 (2008)
Sipser, M.: Halting space bounded computations. Theoret. Comput. Sci. 10, 335–338 (1980)
Sudborough, I.: Efficient algorithms for path system problems and applications to alternating and time-space complexity classes. In: Proceedings of the IEEE Symposium on Foundations of Computer Science, pp. 62–73 (1980)
Szelepcsényi, R.: The method of forced enumeration for nondeterministic automata. Acta Inform. 26, 279–284 (1988)
Szepietowski, A.: Turing Machines with Sublogarithmic Space. LNCS, vol. 843. Springer, Heidelberg (1994)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Geffert, V. (2016). Alternating Demon Space Is Closed Under Complement and Other Simulations for Sublogarithmic Space. In: Brlek, S., Reutenauer, C. (eds) Developments in Language Theory. DLT 2016. Lecture Notes in Computer Science(), vol 9840. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-662-53132-7_16
Download citation
DOI: https://doi.org/10.1007/978-3-662-53132-7_16
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-662-53131-0
Online ISBN: 978-3-662-53132-7
eBook Packages: Computer ScienceComputer Science (R0)