Computer Science and Inf ormation T echnologies V ol. 7, No. 2, July 2026, pp. 196 202 ISSN: 2722-3221, DOI: 10.11591/csit.v7i2.p196-202 196 Complexity of nite state T uring machine with other domain Rajesh K umar 1 , Anju J ain 2 , Rak esh K umar 3 1 Department of Computer Science and Applications, Chhaju Ram Memorial Jat Colle ge, Hisar , India 2 Department of Computer Science, Go v ernment Colle ge Hansi, Hansi, India 3 Department of Computer Science and Applications, K urukshetra Uni v ersity , K urukshetra, India Article Inf o Article history: Recei v ed Sep 21, 2025 Re vised Apr 25, 2026 Accepted May 16, 2026 K eyw ords: Comple xity Finite state automata Non-deterministic FSA P artial function T uring machine ABSTRA CT In this paper , the authors in v estig ate and discussed the non-det erministic state comple xity of certain operations on nite state T uring machine on other domain which includes parti al function and natural function o v er an alphabet se t Σ . It is found that in some boolean operations on said domains, the state comple xity reaches up to upper bound O ( n !) . This result is complement for the operation on Kleen star -free unary and recursi v e languages a ccepted by the nite state T uring machine. This is an open access article under the CC BY -SA license . Corresponding A uthor: Rajesh K umar Department of Computer Sciences and Applications, Chhaju Ram Memorial Jat Colle ge Hisar , Haryana, India Email: rajtaya@kuk.ac.in 1. INTR ODUCTION The descripti v e comple xity issues about operation problem for re gular languages of nite state T uring machine ha v e been discussed by the authors in this paper . The operation problem on a language f amily is dened in term of total states in relation to accepting state by deterministic nite state T uring machine. It is well kno wn that deterministic and non-deterministic nite state machines are equi v alent in terms of their computational po wer [1], [2]. In terms of states, a gi v en n-state NF A, one can al w ays construct an equi v alent DF A with at most 2 n state b ut the space comple xity of this ne wly constructed DF A will be huge. It has been found that in most of the cases when an operation is e xpensi v e [3] for NF A, it is v ery cheap for DF A and vice v ersa. In this paper , the author gi v es tw o e xamples: - In the tw o languages accepted by M- and N-state DF A in Figure 1, after concatenation it has a upper bound of m 2 n t 2 n 1 states [4], where t is the number of nal states it goes m + n + 1 when considering NF As. - When complement operation appl ied to a language, a n-state NF A in Figure 2 gi v es an upper bound of 2 n states [5] whereas by an n-state DF A results in e xactly the same number of states in the complemented DF A and vice v ersa. J ournal homepage: http://iaesprime .com/inde x.php/csit Evaluation Warning : The document was created with Spire.PDF for Python.
Comput Sci Inf T echnol ISSN: 2722-3221 197 q 1 start q 2 q 3 q m 1 0 1 0 1 0 1 0 1 0 0 0 Figure 1. L 1 = { (0 , 1) 1 m | (0,1) Σ } q 1 start q 2 q 3 q m 1 0 1 0 1 0 1 0 1 0 0 0 Figure 2. L 1 = { Σ (0 , 1) 1 m | (0 , 1) Σ } 2. TURING MA CHINE: A ST AND ARD MODEL The e v olution of machine or automata are as: i) nite memory (encloded in state) —DF As/ NF As; ii) unbounded stack —NPD As/DPD As; and iii) unbound memory —TMs. 2.1. Denition A turing machine: T= Q, Σ , τ | {z } Σ τ { } , δ B , Q o , Q F Q, |{z} bl ank sy mbol where δ B = δ B : Q × τ |{z} par tial f unction Q × τ × { L, R } f ( Q × τ ) × ( Q × τ × { L, R } ) . Figure 3 depicts the interaction between these components. Innite tape [6], [7] F S C q [8], [9] Figure 3. F S C q - is the read/ write head of nite state control whereas innite tape- is used for storing input, as scratch pad and also (in the case of T uring machine as transducers) for storing the output 2.1.1. Instantaneous Description (ID) An instantaneous description [10], [11] of a TM is a sequence of the form xq y where x,y τ such that x ̸ = ̸ = y . - There is non-blank symbol either to the left of x or to the right of y , - The rst symbol of x and the last symbol of y are both non-blank, - The rst symbol of y is the symbol immediately under the read/write head, - q is the current state of the FSC. Comple xity of nite state T uring mac hine with other domain (Rajesh K umar) Evaluation Warning : The document was created with Spire.PDF for Python.
198 ISSN: 2722-3221 An ID is also called a conguration. A mo ve of the TM from an ID x 1 q 1 y 1 to another ID x 2 q 2 y 2 is denoted x 1 q 1 y 1 x 2 q 2 y 2 where x 1 = a 1 ....a k ; y 1 = b 1 ....b l is possible pro vided one of the follo wing holds. - δ ( q 1 , b 1 ) = ( q 2 , c, R ) and x 2 = a 1 ....a k c ; y 2 = b 2 ....b l or - δ ( q 1 , b 1 ) = ( q 2 , c, L ) and x 2 = a 1 ....a k 1 ; y 2 = a k cb 1 ....b l . The language accepted by a turing machine T is L ( T )= { x Σ | q 0 x y q f z , y , z τ , q f F } 3. TURING MA CHINE AS A CCEPT ORS Example 1: L = { a n b n | n 1 } [12], [13]. T able 1 describes the transition function of the T uring machine for Example 1. T able 1. T ransition table for turing machine which accept the language L = { a n b n | n 1 } Q τ a b A B q 0 ( q a , A , R ) ( q 2 , B , R ) q a ( q a , A , R ) ( q b , B , R ) ( q a , B , R ) q b ( q b , A , L ) ( q 0 , A , R ) ( q b , B , L ) q 0 ( q 0 , B , R ) ( q f , , R ) q f In an y ID the tape contents are of the follo wing in v ariant property can be used to determine the transition: A a B b (1) if n 0 then the follo wing transi tion could also be included: δ ( q 0 , ) = ( q f , , R ) (2) 3.1. Algorithm Problem: a language L = { xx | x ( a, b ) } . Finding the midpoint of the sub-string and ins erting a symbol C at the midpoi nt. Solution schema: - Starting from the leftmost symbol mo v e to the rightmost symbol and replace the rst by C ”. - Change the leftmost lo wercase symbol to UPPERCASE and the rightmost lo wercase symbol to UPPER- CASE. - Exchange the C with the ne w uppercase s y m bol immediately preceding it. - The tape in v ariant is no w something lik e ( A + B ) . ( a + b ) .C . ( A + B ) . - Repeat steps 2-3 till the tape contents become ( A + B ) C ( A + B ) . - No w change all uppercase letter e xcept C to lo wercase (if necessary). 4. RESUL TS AND DISCUSSION Clearly the T uri ng machine discussed in pre vious section can only compute functions o v er strings o v er a nite alphabet [14], [15]. What about computing functions o v er arbitrary(computable) domains such as the naturals [16], [17], inte gers or the rationals or e v en Cartesian products [18], [19] of such domains or e v en more generally functions from one domain to another? 4.1. Computability in other domains Clearly if the elements of such domains can be represented as string o v er an alphabet, then one can talk about T uring machine computability o v er such domains [20], [21]. Comput Sci Inf T echnol, V ol. 7, No. 2, July 2026: 196–202 Evaluation Warning : The document was created with Spire.PDF for Python.
Comput Sci Inf T echnol ISSN: 2722-3221 199 4.1.1. Denition A set S is represented by elements of another set R if there e xists a partial surjecti v e function [22] R S called the interpretation [23], [24] of R onto S , as in Figure 4, where: - P artial is not e v ery element in R need ha v e a meaning in S . - Surjecti v e is e v ery element of S should ha v e a representation in R . - On the other hand tw o or more elements of R may be interpreted as representing the same element in S . R S Figure 4. Mapping diagram of partial surjecti v e function R S Example: consider the unary representation of the naturals as strings of ‘1 S terminated by a single ‘0‘ . Clearly I : { 0 , 1 } N is a partial function because strings such as ‘00100‘ ha v e no interpretation in N . I is also surjecti v e since e v ery natural does ha v e a representation in { 0 , 1 } . If leading 0 S are allo wed in the representation, then each natural has more than one possible representation. Gi ving tw o sets S 1 and S 2 represented by R 1 and R 2 respecti v ely through interpretations I 1 and I 2 respecti v ely , as in Figure 5. R 1 R 2 S 1 S 2 F I 2 I 1 G Figure 5. A partial function F : R 1 R 2 represents a partial function G : S 1 S 2 if for all r 1 D o m ( I 1 ) , G ( I 1 ( r 1 )) = I 2 ( F ( r 1 )) i.e. the abo v e commutati v e diagram holds Alternati v ely , one may attempt to construct F as a representation of G as a total function, rather than e xpressing the diagram as a partial function. As in Figure 6, gi v en I 1 : R 1 S 1 consider: I 1 1 : S 1 N R { ϕ } where I 1 1 ( s 1 ) = { r 1 R|I 1 ( r 1 ) = s 1 } i.e. I 1 1 for an y s 1 S 1 yield the set of all possible representati on s of s 1 . Then F : R 1 : R 2 may also be e xtended to subsets of R 1 as: F : N R 1 N R 2 is dened for each X 1 R 1 as f ( X 1 ) = { f ( r 1 ) | r 1 X 1 } = X 1 R 2 . But one should be interested in ensuring that the follo wing diagram holds for an y s 1 S 1 , if G ( s 1 ) S 2 , require that: - There is at least one representation r 1 R 1 such that I 2 ( F ( r 1 )) = G S 1 . - r 1 ̸ = r 1 I 1 ( r 1 ) = I 1 ( r 2 ) = s 1 F ( r 1 ) , F ( r 1 ) R 2 I 2 ( F ( r 1 )) = I 2 ( F ( r 1 )) . - There may be representation of s 1 in R 1 for which F ( r 1 ) may not be dened. Comple xity of nite state T uring mac hine with other domain (Rajesh K umar) Evaluation Warning : The document was created with Spire.PDF for Python.
200 ISSN: 2722-3221 S 1 N R 1 S 2 N R 2 I 1 1 F G I 1 2 Figure 6. A function F : N R 1 1 N R 2 2 represents the possibility that F may not be dened for some v alue in f ( X 1 ) = { f ( r 1 ) | r 1 X 1 } = X 1 R 2 4.2. T uring-computable As in Figure 7, a partial function G : S 1 S 2 is T uring-computable [25], [26] if there e xists a T uring-computable function F : Σ Σ on an alphabet Σ and interpretations I 1 , I 2 with I 1 : Σ S 1 and I 2 : Σ S 2 such that: - F or each s 1 S 1 and F ( s 1 ) = s 2 S 2 , x 1 , x 2 Σ [ I 1 ( x 1 ) = s 1 I 2 ( x 2 ) = s 2 F ( x 1 ) = x 2 y 1 Σ [ x 1 ̸ = y 1 I 1 ( y 1 ) = s 1 F ( y 1 ) / Σ I 2 ( F ( y 1 )) = s 2 ]] - F or each s 1 S 1 , such that G ( s 1 ) / s 2 , x 1 Σ [ I 1 ( x 1 ) = s 1 F ( x 1 ) / Σ I 2 ( F ( x 1 )) / s 2 ]] X 1 X 2 S 1 Y 1 S 2 Y 2 F I 1 I 2 I 1 F G I 2 Figure 7. F or an y partial function G : S 1 S 2 where G ( s 1 ) / S 2 if and only if s 1 / D om ( G ) 5. CONCLUSION A consequences of the abo v e denition is that the implem entation of an y function of an y arity is that of a unary function on Σ for the chosen alphabet. F acts: with addition to natural function a binary function is considered which is dened as : + : N × N N . Ho we v er for an y pair of naturals m, n N it can be encoded o v er Σ = { 0 , 1 } in unary with ‘0‘ as a separator between the components of the pair . Hence Σ N × N is dened as I 1 (1 m 01 n ) = ( m, n ) and all other patterns of strings x Σ L (1 01 ) , I 1 ( x ) is undened. Similarly I 2 for all the result is simply dened as: I 2 ( y ) = ( P if y = 1 p 1 undened otherwise W ith abo v e representation the partial function ˆ + : Σ Σ that we require is gi v en by: ˆ +( x ) = ( 1 m + n if x = 1 m 01 n for all m, n 0 undened otherwise An y T uring machine T ˆ + which implement ˆ + is a correct implementation of addition on the naturals. Comput Sci Inf T echnol, V ol. 7, No. 2, July 2026: 196–202 Evaluation Warning : The document was created with Spire.PDF for Python.
Comput Sci Inf T echnol ISSN: 2722-3221 201 FUNDING INFORMA TION Authors state no funding in v olv ed. A UTHOR CONTRIB UTIONS ST A TEMENT This journal uses the Contrib utor Roles T axonomy (CRediT) to recognize indi vidual author contrib utions, reduce authorship disputes, and f acilitate collaboration. Name of A uthor C M So V a F o I R D O E V i Su P Fu Rajesh K umar Anju Jain Rak esh K umar C : C onceptualization I : I n v estig ation V i : V i sualization M : M ethodology R : R esources Su : Su pervision So : So ftw are D : D ata Curation P : P roject Administration V a : V a lidation O : Writing - O riginal Draft Fu : Fu nding Acquisition F o : F o rmal Analysis E : Writing - Re vie w & E diting CONFLICT OF INTEREST ST A TEMENT Authors state no conict of interest. D A T A A V AILABILITY The data that support the ndings of this study are a v ailable from the corresponding author , [RK], upon reasonable request. REFERENCES [1] H. M. Y ao and L. Jiang, “Machine-learning-based PML for the FDTD method, IEEE Antennas and W ire less Propag ation Letters , v ol. 18, no. 1, pp. 192–196, 2019, doi: 10.1109/LA WP .2018.2885570. [2] N. Boull ´ e and A. T o wnsend, “Learning elliptic partial dif ferential equations with randomized linear algebra, F oundations of Computational Mathematics , v ol. 23, no. 2, pp. 709–739, Apr . 2023, doi: 10.1007/s10208-022-09556-w . [3] J. L. Hennessy and D. P atterson, A Ne w Golden Age for Computer Architecture Inno v ations lik e domain-specic hardw are, enhanced security , open instruction sets, and agile chip de v elopment will lead the w ay . Cacm.acm.or g. Accessed Jun. 21, 2025. [Online]. A v ailable: https://cacm.acm.or g/research/a-ne w-golden-age-for -computer -architecture/ [4] G. Jir ´ ask o v ´ a, A. Szabari, and J. ˇ Sebej, “The comple xity of languages resulting from the concatenation operation, in International Conference on Descriptional Comple xity of F ormal Systems , 2016, pp. 153–167, doi: 10.1007/978-3-319-41114-9 12. [5] G. Jir ´ ask o v ´ a and A. Okhotin, “State com ple xity of unambiguous operations on nite automata, Theor etical Computer Science , v ol. 798, pp. 52–64, Dec. 2019, doi: 10.1016/j.tcs.2019.04.008. [6] Y . Oktar , A computing m achinery using a continuous memory tape, arXi v: 2401.02420, 2023. [7] M. Boja ´ nczyk and R. Stef a ´ nski, “Single-use automata and transducers for innite alphabets, in 47th International C olloquium on Automata, Languages, and Programming (ICALP 2020) , 2020, pp. 1–14, doi: 10.4230/LIPIcs.ICALP .2020.113. [8] H. Boche, A. Grigorescu, R. F . Schaefer , and H. V . Poor , Algorithmic computability of the capacity of additi v e colored Gaussian noise channels, in GLOBECOM 2023 - 2023 IEEE Global Communications Conference , Dec. 2023, pp. 4375–4380, doi: 10.1109/GLOBECOM54140.2023.10436918. [9] L. Dartois, P . Gastin, L. G. Guizouarn, R. Go vind, and S. Krishna, “Re v ersible transducers o v er innite w ords, arXi v: 2406.11488, 2024. [10] M. de Benedetto, “Explication as a three-step procedure: the case of the Church-T uring thesis, European Journal for Philosoph y of Science , v ol. 11, no. 1, p. 21, Mar . 2021, doi: 10.1007/s13194-020-00337-2. [11] O. Goldreich, S. Micali, and A. W igderson, “Ho w to play an y mental g ame , or a completeness theorem for protocols with honest ma jority , in Pro viding Sound F oundations for Cryptograph y: On the W ork of Sha Goldw asser and Silvio Micali , 2019, pp. 218–229, doi: 10.1145/3335741.3335755. [12] H. B. Ax elsen and R. Gl ¨ uck, “On re v ersible T uring machines and their function uni v ersality , Acta Informatica , v ol. 53, no. 5, pp. 509–543, Aug. 2016, doi: 10.1007/s00236-015-0253-y . [13] B. Gonc ¸ alv es, “The T uring test is a thought e xperiment, Minds and Machines , v ol. 33, no. 1, pp. 1–31, Mar . 2023, doi: 10.1007/s11023-022-09616-8. [14] F . Ne v en, T . Schwentick, and V . V ianu, “Finite state machines for strings o v er innite alphabets, A CM T ransactions on Computational Logic (T OCL) , v ol. 5, no. 3, pp. 403–435, Jul. 2004, doi: 10.1145/1013560.1013562. [15] Y . F orster , F . K unze, and M. W uttk e, “V eried programming of T uring machi nes in Coq, in Proceedings of the 9th A CM SIGPLAN International Conference on Certied Programs and Proofs , Jan. 2020, pp. 114–128, doi: 10.1145/3372885.3373816. Comple xity of nite state T uring mac hine with other domain (Rajesh K umar) Evaluation Warning : The document was created with Spire.PDF for Python.
202 ISSN: 2722-3221 [16] T . Ha, V . Harizano v , L. Marshall, and H. W alk er , “Comput ability and denability , in Structure and Randomness in Computability and Set Theory , D. Cenzer , C. Porter , and J. Zapletal, Eds., Sing apore: W orld Scientic, 2021, pp. 285–355. [17] K. Mainzer , “Logical thinking becomes automatic, in Articial Intelligence - When do Machines T ak e Ov er? , K. Mainzer , Ed., Berlin: Springer Berlin Heidelber g, 2020, pp. 15–45, doi: 10.1007/978-3-662-59717-0 3. [18] J . Y ou, R. Y ing, and J. Lesk o v ec, “Design space for graph neural netw orks, in Adv ances in Neural Information Processing Systems , H. Larochelle, M. Ranzato, R. Hadsell, M. F . Balcan, and H. Lin, Eds., Ne w Y ork, NY : Curran Associates, Inc, 2020, pp. 17009–17021. [19] B . Jacquet, F . Jamet, and J. Baratgin, “On the pragmatics of the T uring test, in 2021 International Conference on Information and Digital T echnologies (IDT) , Jun. 2021, pp. 123–130, doi: 10.1109/IDT52577.2021.9497570. [20] W . Si e g, “G ¨ odel’ s philosophical challenge (to T uring), Studia Semiotyczne , v ol. 34, no. 1, pp. 57–80, 2020. [21] V . Brattka and P . Hertling, Handbook of computability and comple xity in analysis . Cham: Springer International Publishing, 2021, doi: 10.1007/978-3-030-59234-9. [22] C . Borlido and B. McLean, “Dif ference–restriction algebras of partial functions with operators: discrete duality and completion, Journal of Algebra , v ol. 604, pp. 760–789, Aug. 2022, doi: 10.1016/j.jalgebra.2022.03.039. [23] J. Mayr et al., “Thermal issues in machi ne tools, CIRP Annals , v ol. 61, no. 2, pp. 771–791, 2012, doi: 10.1016/j.cirp.2012.05.008. [24] C. Knapp, “P artial functions and recursion in uni v alent type theory , arXi v: 2011.00272, 2020. [25] Z. Hou, “T uring machi nes and computability , in Fundamenta ls of Logic and Computation: W ith Practical Automated Reasoning and V erication , Z. Hou, Ed., Cham: Springer International Publishing, 2021, pp. 163–205, doi: 10.1007/978-3-030-87882-5 5. [26] H. Pe tersen, “Some remarks on real-time T uring machines, arXi v: 1902.00975, 2019. BIOGRAPHIES OF A UTHORS Rajesh K umar obtained his B.Sc. De gree, Master’ s De gree (Master of Computer Applications) and Ph.D. from K urukshetra Uni v ersity , K urukshetra. Currently , he is head of the Department of Computer Science and Applications, Chhaju Ram Memorial Jat Colle ge, Hisar , Haryana, India. He is a co-author of three books. His research interests are in genetic algorithm, theory of automata, softw are engineering, articial intelligence, design and analysis of algorithm and Linux administration and k ernel design. He can be contacted at email: rajtaya@kuk.ac.in. Anju J ain holds a Ph.D. in Computer Science and Engineering from Guru Jambheshw ar Uni v ersity of Science and T echnology (GJUS&T), India. She has serv ed in v arious academic roles including lecturer , teaching associate, and assistant professor across academies, colle ges, and uni v ersities. A tw o-time Gold Medalist, Dr . Jain has published se v eral research papers in reputed journals and conference proceedings. She possesses nearly 15 years of teaching e xperienc e, with o v er 7 years i n the Computer Science departm ent in higher education in India. Her broad research interests include machine learning, data analytics, data mining, and e v olutionary algorithms. She can be contacted at email: anjuaryan2012@gmail.com. Rak esh K umar obtained his B.Sc. De gree, Master’ s De gree Gold Medalist (Master of Computer Applications) and Ph.D. (Computer Science and Applications) from K urukshetra Uni v ersity , K urukshetra. Currently , he is dean academic af f ai rs, professor and chairperson in the Department of Computer Science and Applications, K urukshetra Uni v ersity , K urukshetra, Haryana, India. His research interests are in genetic algorithm, softw are testing, articial intelligence, and netw orking. He is a senior member of International Association of Computer Science and Information T echnology (IA CSIT). He ca n be contacted at email: rak eshkumar@kuk.ac.in. Comput Sci Inf T echnol, V ol. 7, No. 2, July 2026: 196–202 Evaluation Warning : The document was created with Spire.PDF for Python.