Abstract
For multiparticle finite-state action evolutions, we prove that the observation \( \sigma \)-field admits a resolution involving a third noise which is generated by a random variable with uniform law. The Rees decomposition from semigroup theory and the theory of infinite convolutions are utilized in our proofs.
Similar content being viewed by others
Data Availability
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.
References
Brossard, J., Leuridan, C.: Chaînes de Markov constructives indexées par \(\textbf{Z}\). Ann. Probab. 35(2), 715–731 (2007)
Chaumont, L., Yor, M.: Exercises in Probability. A Guided Tour from Measure Theory to Random Processes, via Conditioning. Cambridge Series in Statistical and Probabilistic Mathematics, 2nd edn. Cambridge University Press, Cambridge (2012)
Cirel’son, B.S.: An example of a stochastic differential equation that has no strong solution. Teor. Verojatnost. i Primenen. 20(2), 427–430 (1975)
Collins, H.S.: Convergence of convolution iterates of measures. Duke Math. J. 29, 259–264 (1962)
Hirayama, T., Yano, K.: Extremal solutions for stochastic equations indexed by negative integers and taking values in compact groups. Stoch. Process. Appl. 120(8), 1404–1423 (2010)
Högnäs, G., Mukherjea, A.: Probability Measures on Semigroups. Convolution Products, Random walks, and Random Matrices. Probability and its Applications (New York), 2nd edn. Springer, New York (2011)
Ito, Y., Sera, T., Yano, K.: Examples of third noise problems for action evolutions with infinite past. In: RIMS Kōkyūroku 2176, Proceedings of Research on the Theory of Random Dynamical Systems and Fractal Geometry in 2019, pp. 20–27 (2021)
Rosenblatt, M.: Limits of convolution sequences of measures on a compact topological semigroup. J. Math. Mech. 9, 293–305 (1960)
Schwarz, Š: Convolution semigroup of measures on compact noncommutative semigroups. Czechoslov. Math. J. 14(89), 95–115 (1964)
Trahtman, A.N.: The road coloring problem. Israel J. Math. 172, 51–60 (2009)
van Handel, R.: On the exchange of intersection and supremum of \(\sigma \)-fields in filtering theory. Israel J. Math. 192(2), 763–784 (2012)
Yano, K.: Random walk in a finite directed graph subject to a road coloring. J. Theor. Probab. 26(1), 259–283 (2013)
Yano, K.: Infinite convolutions of probability measures on polish semigroups. Probab. Surv. 19, 129–159 (2022)
Yano, K., Yor, M.: Around Tsirelson’s equation, or: the evolution process may not explain everything. Probab. Surv. 12, 1–12 (2015)
Yor, M.: Tsirel’son’s equation in discrete time. Probab. Theory Relat. Fields 91(2), 135–152 (1992)
Acknowledgements
The authors would like to thank the referee for a lot of valuable comments which helped improve the earlier versions of this paper. In particular, the two appendices are mainly due to the referee.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors have no competing interests to declare that are relevant to the content of this article.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This research was supported by RIMS and by ISM. Yu Ito: The research of this author was supported by JSPS KAKENHI Grant Number JP18K13431. Toru Sera: Research Fellow of Japan Society for the Promotion of Science. The research of this author was supported by JSPS KAKENHI Grant Numbers JP19J11798 and JP21J00015. Kouji Yano: The research of this author was supported by JSPS KAKENHI Grant Nos JP19H01791, JP19K21834 and JP18K03441 and by JSPS Open Partnership Joint Research Projects Grant No. JPJSBP120209921.
Appendices
Appendix A: The Semigroup Consisting of Mappings
Proof of Proposition 1.11
(i) Let us define K by (1.18) and prove that K is a minimal ideal of S. For \( f \in K \) and \( g,h \in S \), we have \( m_S \le \#(gfhV) \le \#(fV) = m_S \), which shows that K is an ideal of S. Let \( I \subset K \) be an ideal of S. Let \( f \in K \) and \( g \in I \). Since \( fgf|_{fV} \) is a permutation of fV, there exists an integer \( q \ge 1 \) such that \( (fgf)^q \) is identity on fV, which implies \( (fgf)^q f = f \). Hence,
which shows \( I=K \), and thus K is the kernel of S.
(ii) This is obvious.
(iii) Let e be a primitive idempotent of S. Since K is completely simple by Proposition 1.6, we may take \( f \in E(K) \). Then, \( efe \in SKS \subset K \). Since \( efe|_{efeV} \) is a permutation of efeV, there exists an integer \( q \ge 1 \) such that \( (efe)^q \) is identity on efeV, which yields \( (efe)^{q+1} = efe \). If we write \( g := (efe)^{2q} \), we obtain \( eg=ge=g \in E(K) \), which implies \( g=e \) by primitivity. Thus, we obtain \( e \in E(K) \). The converse is obvious since all idempotents of K are primitive.
(iv) Let \( f \in Se = LG \) and take \( (x,g) \in L \times G \) such that \( f = xg \). Since \( g^{-1}f = e \) and \( fe = f \), we have [\( fv=fw \) \( \iff \) \( ev=ew \)] for all \( v,w \in V \), which shows \( \pi (f)=\pi (e) \).
Conversely, let \( f \in S \) be such that \( \pi (f)=\pi (e) \). Then, \( \#(fV) = \#(eV) = m_S \), so that \( f \in K \). Let \( f = xgy \) with \( (x,g,y) \in L \times G \times R \). Since \( \pi (y) = \pi (gy) = \pi (ef) = \pi (f) = \pi (e) \) and \( ye = e \), we obtain \( y = e \), so \( f \in Se \).
(v) Let \( f \in eS = GR \) and take \( (g,y) \in G \times R \) such that \( f = gy \). Then, \( fV = efV \subset eV \). Since \( \#(fV) = \#(eV) = m_S \), we have \( fV=eV \).
Conversely, let \( f \in S \) be such that \( fV=eV \). Then, \( \#(fV) = \#(eV) = m_S \), so that \( f \in K \). Take \( (x,g,y) \in L \times G \times R \) such that \( f = xgy \). Note that \( fe = xgye = xg \) and \( x = feg^{-1} = fg^{-1} \). Since \( xV = fg^{-1}V \subset fV = eV \), we have \( xV = eV \). On the one hand, since e is identity on \( xV = eV \), we have \( exv = xv \) for \( v \in V \). On the other hand, since \( ex = e \), we have \( exv = ev \) for \( v \in V \). We now obtain \( x=e \), so \( f \in eS \).
(vi) This is immediate from (iv) and (v), since \( G = Se \cap eS \). \(\square \)
Appendix B: Another Example
Let
Let \( D = \{ (1,0),(-1,0),(0,1),(0,-1) \} \) and set
Note that S is a finite semigroup with respect to the usual matrix product. In fact,
for \( (a_{21},a_{22}) \in D \), etc. We regard an element of S as a map of V into itself with respect to the usual matrix product. Set
and
so that \( s_0,s_1,s_2,g \in S \). Let \( \mu = (\delta _{s_0} + \delta _{s_1} + \delta _{s_2} + \delta _{g})/4 \) be the uniform law on \( \mathcal {S}(\mu ) = \{ s_0,s_1,s_2,g \} \). It is easy to see that \( \mathcal {S}(\mu ) \) generates S, i.e., \( S = \bigcup _{n=1}^{\infty } \{ s_0,s_1,s_2,g \}^n \). Let us apply Propositions 1.8, 1.11 and 1.12.
If we write
then
and \( gV=gA=gB=A \), and hence we see that \( m_\mu = m_S = \#(A) = \#(B) = 4 \).
Set
Since \( \mathcal {S}(\mu ) \subset S_- \), we have \( \mathcal {S}(\mu ^2) = \mathcal {S}(\mu ) \mathcal {S}(\mu ) \subset S_- S_- = S_+ \). Since \( \mathcal {S}(\mu ^2) \) generates \( S_+ \) and contains the identity map, we see that the left or right random walk on \( S_+ \) whose steps have law \( \mu ^2 \) is aperiodic, whereas the random walk on S whose steps have law \( \mu \) is not. Hence, we obtain \( p=2 \), and consequently the sequence \( \{ \mu ^{2n} \}_{n=1}^{\infty } \) converges to \( \eta \).
Let \( K = \mathcal {S}(\nu ) \) and \( K_+ = \mathcal {S}(\eta ) \) denote the kernels of S and \( S_+ \), respectively. Then,
Set
which is an idempotent of \( S_+ \). Since \( \#(eV) = 4 \), we have \( e \in K_+ \). Let us determine the Rees decompositions \( K=LGR \) and \( K_+=LHR \) at \( e \in E(K_+) \). Set
Then, we have
We now see that we may choose \( \gamma = g \), so that \( C = \{ e,g \} \) and \( G = CH \). We have the following multiplication tables (the table of ab for a and b):
Note that \( -f = fg(-g) \) and \( -k = g(-g)k \).
Let us compute \( \eta ^L \). Let \( \eta ^L = \alpha \delta _e + \beta \delta _f \) for some \( \alpha ,\beta >0 \) with \( \alpha + \beta = 1 \). Note that
On the one hand, we have
On the other hand, by (7.17), we have
Since \( -f = fg(-g) \) and \( -g \in H \), we have \( \delta _{-f} * \omega _H = \delta _{fg} * \omega _H \), so
Hence, we obtain \( \alpha = 2/3 \) and \( \beta = 1/3 \), so that
By a similar argument, using the identities \( -k = g(-g)k \), \( \omega _H * \delta _g = \delta _g * \omega _H \) and
we obtain
Note that \( eV = \{ v_1,v_2,v_3,v_4 \} \) and \( fV = \{ fv_1,fv_2,fv_3,fv_4 \} \), where
By (3.1) and (3.2), we see that
and
We have the following multiplication table (the table of sv for s and v):
From this table, we see that we may take a set W as
which is a minimal subset of \( W_\mu \) such that \( eW_\mu = GW \).
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Ito, Y., Sera, T. & Yano, K. Resolution of Sigma-Fields for Multiparticle Finite-State Action Evolutions with Infinite Past. J Theor Probab 36, 1368–1399 (2023). https://doi.org/10.1007/s10959-022-01219-4
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10959-022-01219-4
Keywords
- Evolution process
- Tsirelson’s equation
- Resolution of sigma-fields
- Third noise
- Iteration of random mappings