All articles published by MDPI are made immediately available worldwide under an open access license. No special
permission is required to reuse all or part of the article published by MDPI, including figures and tables. For
articles published under an open access Creative Common CC BY license, any part of the article may be reused without
permission provided that the original article is clearly cited. For more information, please refer to
Feature papers represent the most advanced research with significant potential for high impact in the field. A Feature
Paper should be a substantial original Article that involves several techniques or approaches, provides an outlook for
future research directions and describes possible research applications.
Feature papers are submitted upon individual invitation or recommendation by the scientific editors and must receive
positive feedback from the reviewers.
Editor’s Choice articles are based on recommendations by the scientific editors of MDPI journals from around the world.
Editors select a small number of articles recently published in the journal that they believe will be particularly
interesting to readers, or important in the respective research area. The aim is to provide a snapshot of some of the
most exciting work published in the various research areas of the journal.
In the paper we define three kinds of entropy of a fuzzy dynamical system using different entropies of fuzzy partitions. It is shown that different definitions of the entropy of fuzzy partitions lead to different notions of entropies of fuzzy dynamical systems. The relationships between these entropies are studied and connections with the classical case are mentioned as well. Finally, an analogy of the Kolmogorov–Sinai Theorem on generators is proved for fuzzy dynamical systems.
The notion of entropy is basic in information theory [1,2]; it is suitable for measuring the value of information which we get from a realization of the considered experiment. A customary mathematical model of a random experiment in the information theory is a measurable partition of a probability space. Partitions are standardly defined within classical, crisp sets. It turned out however, that for solving real problems partitions defined within the concept of fuzzy sets [3,4] are more suitable. That was a reason why several concepts of generalization of the classical set partition to a fuzzy partition [5,6,7,8,9,10] have been created. A fuzzy partition can serve as a mathematical model of the random experiment whose results are vaguely defined events, the so-called fuzzy events. Kolmogorov and Sinai  (see also ) used the entropy to prove the existence of non-isomorphic Bernoulli shifts (Example 1). Because the Kolmogorov and Sinai theory of entropy of classical dynamical systems has many important and interesting applications, it is reasonable to also expect similar results in the fuzzy case.
In this paper we present our results concerning the entropy of fuzzy dynamical system based on a given probability space. The results represent fuzzy generalizations of some concepts from the classical probability theory. First, we briefly repeat some basic facts from the theory of fuzzy partitions (Section 2) and the classical Kolmogorov–Sinai theory (Section 3). The presented concepts of entropy of fuzzy partitions (Riečan–Dumitrescu, Maličky, and Hudetz entropy) were used to define three kinds of entropy of a fuzzy dynamical system (Section 4). We study the relationships between these entropies and also connections with the classical case. We obtain the measure which can distinguish non-isomorphic dynamical systems more sensitively than the Kolmogorov–Sinai entropy (Theorem 4). Finally, we prove an analogy of the Kolmogorov–Sinai Theorem on generators for the case of fuzzy dynamical systems. The final section presents conclusions and some suggestions for further research. It is noted that certain basic studies on entropy of fuzzy partitions and related notions were done in [13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31].
2. Fuzzy Partitions
In our considerations the Kolmogorov name appears twice. First the Shannon entropy has been used for the distiguishing non-isomorphic dynamical systems by the Kolmogorov–Sinai entropy. We generalize the distinguishing to the fuzzy case. Secondly the whole modern probability theory and mathematical statistics with applications is based on the set theory, and this method was suggested by Kolmogorov. The main prerequisite of the Kolmogorov approach (cf. ) is the identification of the notion of an event with the notion of a set. So consider a non-empty set , some subsets of will be called events. Denote by the family of all events. In the probability theory it is assumed that is a algebra.
Definition 1.A familyof subsets of a non-empty setis called aalgebraif the following conditions are satisfied:
The couplewill be called a measurable space.
Definition 2.Letbe a measurable space. A mappingis called a probability measure if the following properties are satisfied:
The tripletis called a probability space.
If we have a set and then we have only two possibilities: or The set A can be characterized by the characteristic function On the other hand a fuzzy set is a mapping . Analogously to the algebra of sets, we consider a tribe of fuzzy sets.
Definition 3.By a tribe of fuzzy subsets of a setwe shall mean a family F of functionssatisfying the following conditions:
if , then ,
The elements of F are called fuzzy events. If is a algebra, then is a tribe. Another example of a tribe is the family F of all functions measurable with respect to Analogously to the notion of a probability on a algebra we introduce the notion of a state m on F, m :
Definition 4.Let F be a tribe. By a state on F we mean a mapping m :satisfying the following conditions:
if F, then
if F then
One of the nicest results in the theory is the Butnariu and Klement representation theorem  (see also Theorem 8.1.12 in ).
Theorem 1.Let F be a tribe and m : F be a state. Then there exists a probability measure P such that
for every F.
Recall that P is defined on the algebra T = }. Hence it is reasonable to consider a probability space and the family F of all S-measurable functions . The following concept was used, for example, in [6,23].
Definition 5.Letbe a probability space, F be the family of all S-measurable functions(i.e.,). By a fuzzy partition (more precisely F—partition) we understand any sequencesuch that:
Evidently, if 𝒜 = = are fuzzy partitions of then the system 𝒜 : is also a fuzzy partition of We put ... A usual measurable partition of (i.e., each finite sequence of measurable subsets of such that and ()) can be regarded as a fuzzy partition, if we consider instead of Indeed:
3. Kolmogorov–Sinai Entropy
An inspiration for fuzzy entropy was the entropy of the classical partition.
Definition 6.Letbe a probability space, 𝒜 = be an S-measurable partition of. Then the Kolmogorov–Sinai entropy of 𝒜 is the number:
whereis the Shannon entropy defined via:
If 𝒜, are two partitions of then . The symbol has a similar meaning.
Of course, the most important application of Kolmogorov–Sinai entropy has occured in dynamical systems.
Definition 7.By a dynamical system we mean the quadruplewhereis a probability space andis a measure preserving transformation (i.e., andfor any).
Example 1. Let , , , , be the algebra generated by the family of all subsets of the form , and be the probability generated by the equalities , and the mapping by the equality:
Then is a dynamical system, so-called Bernoulli shift (the independent repetition of the experiment ).
Let be an S-measurable partition of In the following, by the partition is denoted. The partition ... represents an experiment consisting of n realizations 𝒜, ,…, of experiment 𝒜. The entropy of experiment 𝒜 with respect to T is defined via:
Definition 8.The Kolmogorov–Sinai entropy of dynamical systemis defined by the formula
where the supremum is taken over all S-measurable partitions 𝒜 of.
If two dynamical systems are isomorphic, then they have the same entropy. It solves the existence of non-isomorphic Bernoulli shifts. Probably one of the most important results of the theory of invariant measures for practical purposes is the Kolmogorov–Sinai Theorem stating that h(T) = 𝒜), whenever A is a partition generating the given algebra S (i.e., a measurable partition such that S). In the following section, we give an analogy of this theorem for the case of fuzzy dynamical systems.
4. The Entropy of Fuzzy Dynamical Systems
Let us return to the fuzzy case. Let a probability space be given. Each fuzzy partition 𝒜 = of represents in the sense of the classical probability theory a random experiment with a finite number of outcomes (which are fuzzy events) with a probability distribution since for and This is a motivation for the following definition.
Definition 9.Letbe a probability space andbe a fuzzy partition of . PutThen the entropy of 𝒜 isgiven by the formula:
In the preceding section we have defined a dynamical system Now we shall define the fuzzy dynamical system.
Definition 10.Letbe a probability space, F be the family of all S-measurable functionsf: Then the quadruple , F, m, , where F F is m-invariant (i.e.,for all F), is called a fuzzy dynamical system.
Example 2. Let be a measure P preserving map. Define F F by the formula:
hence is invariant.
Example 3. Let be a classical dynamical system. Put F Then the system , F, m, , where F F is defined by (1), is a fuzzy dynamical system. By this procedure the classical model can be embedded to a fuzzy one.
Lemma 1.Let F, be a fuzzy dynamical system, 𝒜 be a fuzzy partition ofThen the following limit exists:
Then for any n, , and this inequality implies the existence of . ☐
Definition 11.Let , F, m, be a fuzzy dynamical system. For any non-emptydefine the Riečan–Dumitrescu entropyof , F, m, by the equality:
where the supremum is taken over all fuzzy partitions 𝒜 .
From the following example it follows that the entropy is a fuzzy generalization of the Kolmogorov–Sinai entropy.
Example 4. Let be a dynamical system. Put and define F F by (1). Then is the Kolmogorov–Sinai entropy.
The main result in the Riečan–Dumitrescu entropy is the following theorem on generators (cf. ).
Theorem 2.Letbe an S-measurable partition ofgenerating S, i.e.,Then, for any fuzzy partitionthe following inequality holds:
Of course, if contains all constant functions, then . This defect can be removed by two other constructions, by means of the Maličky entropy and the Hudetz entropy.
In the Riečan–Dumitrescu definition we considered the entropy:
for any fuzzy partition 𝒜. Instead of this number, we will use the number 𝒜, ), …,)) defined as follows. If 𝒜, are two fuzzy partitions, 𝒜 = = then we write 𝒜 if there is a partition of the set such that for any .
Definition 12.Let 𝒜 be a fuzzy partition. Then we define
It is noted that this approach was suggested by Maličky and Riečan in , but only for the case of classical dynamical systems. The above definition includes a more general case. Similarly as in Lemma 1, the following assertion can be proved.
Lemma 2.Let 𝒜 be any fuzzy partition of. Then the following limit exists:
Therefore we are able to define the entropy of a fuzzy dynamical system (, F, m, ).
Definition 13.Let F. Then the entropyof , F, m, is defined by the equality:
where the supremum is taken over all fuzzy partitions.
Now we can compare the entropy with the Riečan–Dumitrescu entropy.
Theorem 3.For any F it holds:
Proof. Let 𝒜 be a fuzzy partition, , . Then:
for any . Therefore ☐
Theorem 4.Letbe a dynamical system. LetandThen:
Proof. Let 𝒜 be an S-partition. Then:
Note that the oposite inequality is proved for some in .
Let be a dynamical system. In the following we shall consider a fuzzy dynamical system , F, m, , where the mapping F F is defined by the formula We shall consider the entropy suggested and studied by Hudetz in [36,37,38].
Definition 14.Let 𝒜 =be afuzzy partition of. Then the Hudetz entropy of 𝒜 is defined by the equality:
Using the Hudetz entropy of fuzzy partition we will define the entropy of fuzzy dynamical systems. The possibility of this definition is based on the following theorem.
Theorem 5.Let , F, m, be a fuzzy dynamical system, 𝒜 be a fuzzypartitionofThen the following limit exists:
Proof. Let 𝒜 be a fuzzy partition. Since we get:
By the principle of mathematical induction we get:
and therefore, )) exists.
Moreover, we have:
Definition 15.For any non-empty F define the entropyof a fuzzy dynamical system , F, m, by the equality:
where the supremum is taken over all F-partitions.
The following theorem is a fuzzy analogy of Kolmogorov–Sinai Theorem on generators.
Theorem 6.Letbe an S-measurable partition ofgenerating S such that F. Then:
Proof. Let 𝒜 be a fuzzy partition of It is sufficient to prove the inequality . Based on Theorem 2 we have:
According to Theorem 5:
By the combination of Equations (2) and (3) we get that
In this paper we study the entropy of fuzzy partitions and the entropy of fuzzy dynamical systems. The presented concepts of entropy of fuzzy partitions were used to define three kinds of entropy of a fuzzy dynamical system. The relationships between these entropies are studied. The presented measures can be considered as measures of information of experiments whose outcomes are vaguely defined events, the so-called fuzzy events. Finally, we prove an analogy of the Kolmogorov–Sinai Theorem on generators for the case of fuzzy dynamical systems.
Similarly to the set theory the fuzzy set theory has also been shown to be useful in many applications of mathematics as well as in the theoretical research. We hope that also the present text can be presented as an illustration of the fact. Of course, there exists a remarkable generalization of fuzzy set theory. It was suggested by K. Atanassov and it is named IF-set theory [39,40]. Instead of one fuzzy set f: IF-set is a pair of fuzzy sets such that The function is called the membership function, the function the non-membership function. If we have a fuzzy set then it can be represented as an IF-set It was reasonable to construct the probability theory on families of IF-sets (see e.g., [34,41]).
There are some results about the entropy on IF-sets (cf. ). Namely, any IF-set can be embedded to a suitable MV–algebra (multivalued algebra). MV-algebras (cf. [43,44,45]) play in multi-valued logic similar role as Boolean algebras in two-valued logic. There are at least two ways for further research in the area. The first way: to study the IF-entropy without using MV-algebras, and by this way to achieve some applications. The second way: to study the entropy on MV-algebras and some of its generalizations as D-posets (cf. ), effect algebras (cf. ), or A-posets (cf. [48,49,50]).
The authors thank the editor and the referees for their valuable comments and suggestions.
Both authors contributed equally and significantly in writing this article. Both authors have read and approved the final manuscript.
Conflicts of Interest
The authors declare no conflict of interest.
Shannon, C.E. A Mathematical Theory of Communication. Bell Syst. Tech. J.1948, 27, 379–423. [Google Scholar] [CrossRef]
Gray, R.M. Entropy and Information Theory; Springer: Berlin/Heidelberg, Germany, 2009. [Google Scholar]
Dumitrescu, D. Entropy of a fuzzy dynamical system. Fuzzy Sets Syst.1995, 70, 45–57. [Google Scholar] [CrossRef]
Riečan, B. An entropy construction inspired by fuzzy sets. Soft Comput.2003, 7, 486–488. [Google Scholar]
Khare, M. Fuzzy σ-algebras and conditional entropy. Fuzzy Sets Syst.1999, 102, 287–292. [Google Scholar] [CrossRef]
Srivastava, P.; Khare, M.; Srivastava, Y.K. m-Equivalence, entropy and F-dynamical systems. Fuzzy Sets Syst.2001, 121, 275–283. [Google Scholar] [CrossRef]
Khare, M.; Roy, S. Entropy of quantum dynamical systems and sufficient families in orthomodular lattices with Bayessian state. Commun. Theor. Phys.2008, 50, 551–556. [Google Scholar] [CrossRef]
Khare, M.; Roy, S. Conditional entropy and the Rokhlin metric on an orthomodular lattice with Bayessian state. Int. J. Theor. Phys.2008, 47, 1386–1396. [Google Scholar] [CrossRef]
Rahimi, M.; Assari, A.; Ramezani, F. A Local Approach to Yager Entropy of Dynamical Systems. Int. J. Fuzzy Syst.2015, 1. [Google Scholar] [CrossRef]
Riečan, B. On a type of entropy of dynamical systems. Tatra Mt. Math. Publ.1992, 1, 135–140. [Google Scholar]
Riečan, B. On some modifications of the entropy of dynamical systems. Atti del Seminario Matematico e Fisico dell Universita di Modena1994, 42, 157–166. [Google Scholar]
Riečan, B. On the entropy and generators of dynamical systems. Appl. Math.1996, 41, 161–168. [Google Scholar]
Markechová, D. The entropy of fuzzy dynamical systems and generators. Fuzzy Sets Syst.1992, 48, 351–363. [Google Scholar] [CrossRef]
Markechová, D. A note to the Kolmogorov–Sinai entropy of fuzzy dynamical systems. Fuzzy Sets Syst.1994, 64, 87–90. [Google Scholar] [CrossRef]
Riečan, B.; Markechová, D. The Kolmogorov–Sinai theorem on generators for fuzzy dynamical systems. Demonstr. Math.1997, 30, 429–436. [Google Scholar]
Markechová, D. Entropy and mutual information of experiments in the fuzzy case. Neural Netw. World2013, 23, 339–349. [Google Scholar] [CrossRef]
Hu, Q.; Yu, D.; Xie, Z.; Liu, J. Fuzzy probabilistic approximation spaces and their information measures. IEEE Trans. Fuzzy Syst.2006, 14, 191–201. [Google Scholar]
Yu, D.; Hu, Q.; Wu, C. Uncertainty measures for fuzzy relations and their applications. Appl. Soft Comput.2007, 7, 1135–1143. [Google Scholar] [CrossRef]
Kolmogorov, A.N. Foundations of the Theory of Probability; Chelsea Publishing Company: New York, NY, USA, 1950. [Google Scholar]
Butnariu, D.; Klement, E.P. Triangular Norm-Based Measures and Games with Fuzzy Coalitions; Springer: Berlin/Heidelberg, Germany, 1993. [Google Scholar]
Riečan, B. Analysis of Fuzzy Logic Models. In Intelligent Systems; Koleshko, V.M., Ed.; InTech: Rijeka, Croatia, 2012; pp. 219–244. [Google Scholar]
Maličký, P.; Riečan, B. On the entropy of dynamical systems. In Proceedings of the Conference Ergodic Theory and Related Topics II, Georgenthal, Germany, 20–25 April 1986.
Hudetz, T. Space-time dynamical entropy for quantum systems. Lett. Math. Phys.1988, 16, 151–161. [Google Scholar] [CrossRef]
Hudetz, T. Algebraic topological entropy. In Nonlinear Dynamics and Quantum Dynamical Systems; Akademie Verlag: Berlin, Germany, 1990; pp. 27–41. [Google Scholar]
Hudetz, T. Quantum Topological Entropy: First Step of a “Pedestrian“ Approach. In Quantum Probability and Related Topics; World Scientific Publishing Company: Singapore, Singapore, 1993; pp. 237–261. [Google Scholar]
Atanassov, K.T. Intuitionistic Fuzzy Sets: Theory and Applications; Springer: Berlin/Heidelberg, Germany, 1999. [Google Scholar]
Atanassov, K.; Riečan, B. On two operations over intuitionistic fuzzy sets. J. Appl. Math. Stat. Inform.2006, 2, 145–148. [Google Scholar] [CrossRef]
Riečan, B. Probability Theory on IF Events. In Algebraic and Proof-Theoretic Aspects of Non-Classical Logics; Aguzzoli, S., Ciabattoni, A., Gerla, B., Manara, C., Marra, V., Eds.; Springer: Berlin/Heidelberg, Germany, 2007; pp. 290–308. [Google Scholar]
Ďurica, M. Entropy on IF-events. Notes on Intuitionistic Fuzzy Sets2007, 13, 30–40. [Google Scholar]
Di Nola, A.; Dvurečenskij, A. Product MV-algebras. Multiple-Valued Log.2001, 6, 193–215. [Google Scholar]
Riečan, B.; Mundici, D. Probability on MV-algebras. In Handbook of Measure Theory; Pap, E., Ed.; Elsevier: Amsterdam, The Netherlands, 2002; pp. 869–910. [Google Scholar]
Riečan, B. Kolmogorov–Sinaj entropy on MV-algebras. Int. J. Theor. Phys.2005, 44, 1041–1052. [Google Scholar] [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely
those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or
the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas,
methods, instructions or products referred to in the content.