• 沒有找到結果。

Lower Bound of Entropy

在文檔中 Pattern Generation Problems (頁 117-124)

3.4 Connecting Operator

3.4.2 Lower Bound of Entropy

In this subsection, the connecting operator Cˆx;m3;m1m2 is employed to esti-mate the lower bound of entropy and in particular, to verify the positivity of entropy.

Definition 3.22. Let X = (X1,· · · , XM)t, where Xk are N × N matrices.

Define the summation of Xk by

|X| =

Note that, (3.4.42) implies

|Xx;mˆ 1,m2,m3| =

2(m1−1)m2X

k=1

A(k)x;mˆ 1,m2,m3= Aˆx;m1,m2,m3. (3.4.44)

As usual, the set of all matrices with the same order can be partially ordered.

114 Pattern Generation Problems Definition 3.23. Let M = [Mij] and N = [Nij] be two M × M matrices, M≥ N if Mij ≥ Nij for all 1 ≤ i, j ≤ M.

Notably, if A2 ≥ A2 then An ≥ An for all n ≥ 2. Therefore, h(A2) ≥ h(A2). Hence, the spatial entropy as a function of A2 is monotonic with respect to the partial order ≥.

Definition 3.24. A P + 1 multiple index αp ≡ (α1α2· · · αpαP+1) (3.4.45)

is called a periodic cycle if

αP+1 = α1. (3.4.46)

It is called diagonal cycle if (3.4.46) holds and αi ∈ {2m2(i−1)+i|1≤i≤22m2} (3.4.47)

for each 1 ≤ i ≤ P + 1. For a diagonal cycle (3.4.45)

¯

αP = α1; α2; · · · ; αP (3.4.48)

and

¯

αPn= ¯αP; ¯αP; · · · ; ¯αP. (n-times) First, prove the following Lemma.

Lemma 3.25. Let m1 ≥ 2, m2 ≥ 2, P ≥ 1, αP be a diagonal cycle. Then, for any m3 ≥ 1,

ρ(Amxˆ;2×m1 2×(m3P+2)) (3.4.49)

≥ ρ(|(Sx;mˆ 3;m1m21α2Sˆx;m3;m1m22α3· · · Sx;mˆ 3;m1m2PαP+1)m3Xx;mˆ 1,m2,2;α1|).

Proof. Since αP is a periodic cycle, Theorem 3.21 implies Xˆx;m1,m2,m3P+2; ¯αPm1

(3.4.50)

= (Sx;mˆ 3;m1m21α2Sx;mˆ 3;m1m22α3· · · Sˆx;m3;m1m2PαP+1)nXx;mˆ 1,m2,2;α1. (3.4.51)

Furthermore αP is diagonal and |Xx;mˆ 1,m2,m3P+2; ¯αPm1| = Ax;mˆ 1,m2,m3P+2; ¯αPm1

lies in the diagonal part of (3.4.41) with m3+ P = m3P + 2, therefore ρ(Amx;mˆ1 1,m2,m3P+2) ≥ ρ(|Xˆx;m1,m2,m3P+2; ¯αPm1|).

(3.4.52)

Therefore, (3.4.49) follows from (3.4.50) and (3.4.52). The proof is complete.

3.4. CONNECTING OPERATOR 115 The following Lemma is valuable in studying maximum eigenvalue of (3.4.52).

Lemma 3.26. For any m1 ≥ 2, m2 ≥ 2, 1 ≤ k ≤ 2(m1−1)m2 and α1 ∈ {(i − 1)2m2 + i|1 ≤ i ≤ 2m2}, if

tr(A(k)x;mˆ 1,m2,2;α) = 0, (3.4.53)

then for all 1 ≤ ℓ ≤ 2(m1−1)m2,

(Sx;mˆ 3;m1m21α2)kℓ = 0, (3.4.54)

for all α2 ∈ {(i − 1)2m2 + i|1 ≤ i ≤ 2m2}, i.e., the k-th rows of matrices Sˆx;m3;m1m21α2 are zeros. Furthermore, for any diagonal cycle αP, let U =

(u1u2· · · u2m2(m1−1)) be an eigenvector of Sˆx;m3;m1m21α2Sx;mˆ 3;m1m22α3· · · Sˆx;m3;m1m2Pα1, if uk 6= 0 for some 1 ≤ k ≤ 2(m1−1)m2, then tr(A(k)x;mˆ 1,m2,2;α1) > 0.

Proof. Since A(k)x;mˆ 1,m2,2;α1 can be expressed as (3.4.33). Therefore, tr(A(k)x;mˆ 1,m2,2;α1) = 0 if and only if (3.4.54) holds for all 1 ≤ ℓ ≤ 2(m1−1)m2. The second part of

the Lemma follows easily from the first part. The proof is complete.

By Lemma 3.25 and Lemma 3.26, the lower bound of entropy can be obtained as follows.

Theorem 3.27. Let α1α2· · · αPα1be a diagonal cycle. Then for any m1 ≥ 2, m2 ≥ 2,

h(Ax;2×2×2) (3.4.55)

≥ 1

m1m2P log ρ(Sx;mˆ 3;m1m21α2Sx;mˆ 3;m1m22α3· · · Sˆx;m3;m1m2Pα1).

In particular, if a diagonal cycle α1α2· · · αPα1 exists and m1 ≥ 2, m2 ≥ 2

such that ρ(Sx;mˆ 3;m1m21α2Sx;mˆ 3;m1m22α3· · · Sx;mˆ 3;m1m2Pα1) > 1, then h(Ax;2×2×2) >

0.

Proof. First, by the similar method in the proof of Lemma 2.10 and Lemma 2.11 and Theorem 2.12 in [5] we have

lim sup

m3→∞

1

m3(log ρ(|(Sˆx;m3;m1m21α2Sx;mˆ 3;m1m22α3· · · Sˆx;m3;m1m2Pα1)nXx;mˆ 1,m2,2;α1|))

= log ρ(Sx;mˆ 3;m1m21α2Sˆx;m3;m1m22α3· · · Sx;mˆ 3;m1m2Pα1).

(3.4.56) Now, show that h(Ax;2×2×2) ≥ 1

m1m2P lim sup

m3→∞

1

m3(log ρ(|(Sx;mˆ 3;m1m21α2Sx;mˆ 3;m1m22α3· · · Sˆx;m3;m1m2Pα1)nXx;mˆ 1,m2,2;α1|))

116 Pattern Generation Problems Indeed, from (3.3.18) and (3.4.49),

h(Ax;2×2×2) = lim

m2m3→∞

1 (m3P + 2)m2

log ρ(Aˆx;2×m2×(m3P+2))

= lim

m2m3→∞

1

m1(m3P + 2)m2

log ρ(Amxˆ;2×m1 2×(m3P+2))

≥ 1

m1m2P lim sup

m3→∞

1

m3(log ρ(|(Sˆx;m3;m1m21α2Sx;mˆ 3;m1m22α3· · · Sx;mˆ 3;m1m2Pα1)nXˆx;m1,m2,2;α1|)).

And by (3.4.56), the proof is complete.

Example 3.28. Consider

Tx;2×2×2 = ⊗(G ⊗ E)2. Then, it is easy to check that

Cˆx;m3;22;11 =



1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1



.

Therefore,

h(Tx;2×2×2) ≥ log 2 2 .

Moreover, in Proposition3.15it can be shown that h(Tx;2×2×2) = log g where g = 1+25.

Reference

[1] J.C. Ban, K.P. Chien and S.S. Lin, Spatial disorder of CNN-with asym-metric output funtion, International J. of Bifurcation and Chaos 11, (2001) 2085-2095.

[2] J.C. Ban, C.H. Hsu AND S.S.Lin, Spatial disorder of cellular neural network-with biased term, International J. of Bifurcation and Chaos 12, (2002) 525-534.

[3] J.C. Ban, S.S. Lin and C.W. Shih, Exact number of mosaic patterns in one-dimensional cellular neural networks, International J. of Bifurcation and Chaos 11, (2001) 1645-1653.

[4] J.C. Ban and S.S. Lin, Patterns generation and transition matrices in multi-dimensional lattice models, Discrete Contin. Dyn. Syst. 13, (2005), no. 3, 637-658.

[5] J.C. Ban, S.S. Lin and Y.H. Lin, Patterns generation and spatial entropy in two-dimensional lattice models, Asian Journal of Mathematics, to appear.

[6] J.C. Ban, S.S. Lin and Y.H. Lin, Primitivity of subshifts of finite type in two-dimensional lattice models, preprint (2005).

[7] J.C. Ban, S.S. Lin and Y.H. Lin, Patterns Generation and Spatial Entropy in Three-dimensional Lattice Models (II): Trace Operators, preprint (2007).

[8] P.W. Bates and A. Chmaj, A discrete convolution model for phase tran-sitions, Arch. Rat. Mech. Anal 150, (1999) 281-305.

[9] P.W. Bates, K. Lu and B. Wang, Attractors for lattice dynamical sys-tems, International J. of Bifurcation and Chaos 11, (2001) 143-153.

[10] J. Bell, Some threshold results for modes of myelinated nerves, Math.

Biosci. 54, (1981) 181-190.

117

118 REFERENCE [11] J. Bell and C. Cosner, Threshold behavior and propagation for nonlinear differential-difference systems motivated by modeling myelinated axons, Quart. Appl. Math. 42, (1984) 1-14.

[12] R. BELLMAN, Introduction to matrix analysis, Mc Graw-Hill, N.Y.

(1970).

[13] R. J. Baxter, Eight-vertex model in lattice statistics., Phys. Rev. Lett.

26, (1971) 832-833.

[14] J.W. Cahn, Theory of crystal growth and interface motion in crystalline materials, Acta Metallurgica 8, (1960) 554-562.

[15] S.N. Chow and J. Mallet-paret, Pattern formation and spatial chaos in lattice dynamical systems II, IEEE Trans. Circuits Systems 42, (1995) 752-756.

[16] S.N. Chow, J. Mallet-paret and E. S. Van Vleck, Dynamics of lattice dif-ferential equations, International J. of Bifurcation and Chaos 6, (1996) 1605-1621.

[17] S.N. Chow, J. Mallet-paret and E.S. Van Vleck, Pattern formation and spatial chaos in spatially discrete evolution equations, Random Comput.

Dynam. 4, (1996) 109-178.

[18] L.O. Chua, CNN: A paradigm for complexity (World Scientific Series on Nonlinear Science, Series A,31. World Scietific, Singapore, 1998)

[19] L.O. Chua, K.R. Crounse, M. Hasler and P. Thiran, Pattern formation properties of autonomous cellular neural networks, IEEE Trans. Circuits Systems 42, (1995) 757-774.

[20] L.O. Chua and T. Roska, The CNN paradigm, IEEE Trans. Circuits Systems 40, (1993) 147-156.

[21] L.O. Chua and L. Yang, Cellular neural networks: Theory, IEEE Trans.

Circuits Systems 35, (1988) 1257-1272.

[22] L.O. Chua and L. Yang, Cellular neural networks: Applications, IEEE Trans. Circuits Systems 35, (1988) 1273-1290.

[23] G.B. Ermentrout, Stable periodic solutions to discrete and continuum arrays of weakly coupled nonlinear oscillators, SIAM J. Appl. Math. 52, (1992) 1665-1687.

REFERENCE 119 [24] G.B. Ermentrout and N. Kopell, Inhibition-produced patterning in chains of coupled nonlinear oscillators, SIAM J. Appl. Math. 54, (1994) 478-507.

[25] G.B. Ermentrout, N. Kopell and T. L. Williams, On chains of oscillators forced at one end, SIAM J. Appl. Math. 51, (1991) 1397-1417.

[26] T. Eveneux and J.P. Laplante, Propagation failure in arrays of coupled bistable chemical reactors, J. Phys. Chem. 96, (1992) 4931-4934.

[27] W.J. Firth, Optical memory and spatial chaos, Phys. Rev. Lett. 61, (1988) 329-332.

[28] F.R. Gantmacher, The theory of matrices (2 vols. Chelsea, New York, 1959)

[29] R.A. Horn and C.R. Johnson, Matrix analysis (Cambridge University Press, Cambridge, 1990)

[30] C.H. Hsu, J. Juang, S.S. Lin, and W.W. Lin, Cellular neural networks:

local patterns for general template, International J. of Bifurcation and Chaos 10, (2000) 1645-1659.

[31] J. Juang and S.S. Lin, Cellular Neural Networks: Mosaic pattern and spatial chaos, SIAM J. Appl. Math. 60, (2000) 891-915.

[32] J. Juang, S.S. Lin, W.W. Lin and S.F. Shieh, The spatial entropy of two-dimensional subshifts of finite type, International J. of Bifurcation and Chaos 10, (2000) 2845-2852.

[33] J.P. Keener, Propagation and its failure in coupled systems of discrete excitable cells, SIAM J. Appl. Math. 47, (1987) 556-572.

[34] J.P. Keener, The effects of discrete gap junction coupling on propagation in myocardium, J. Theor. Biol. 148, (1991) 49-82.

[35] A.L. Kimball, A. Varghese and R.L. Winslow, Simulating cardiac sinus and atrial netwok dynamics on the connection machine, Phys. D 64, (1993) 281-298.

[36] E.H. Lieb, Exact solution of the problem of the entropy of two-dimensional ice, Phys. Rev. Lett. 18, (1967) 692-694.

[37] E.H. Lieb, Exact solution of the f model of an antiferroelectric, Phys.

Rev. Lett. 18, (1967) 1046-1048.

120 REFERENCE [38] E.H. Lieb, Exact Solution of the Two-Dimensional Slater KDP Model

of a Ferroelectric, Phys. Rev. Lett. 19, (1967) 108-110.

[39] E.H. Lieb, The Residual Entropy of Square Ice, Phys. Rev. 162, (1967) 162-172.

[40] E.H. Lieb, Ice, ferro- and antiferroelectrics, in methods and problems in theoretical physics, in honour of R.E. Peierls, proceedings of the 1967 Birmingham conference, (North-Holland, 1970) 21-28.

[41] S.S. Lin and T.S. Yang, Spatial entropy of one-dimensional celluar neu-ral network, International J. of Bifurcation and Chaos 10, (2000) 2129-2140.

[42] S.S. Lin and T.S. Yang, On the spatial entropy and patterns of two-dimensional cellular neural networks, International J. of Bifurcation and Chaos 12, (2002) 115-128.

[43] D. Lind and B. Marcus, An introduction to symbolic dynamics and cod-ing, Cambridge University Press, Cambridge, (1995).

[44] A. Lindenmayer and P. Prusinkiewicz, The algorithmic beauty of plants, Springer-Verlag, New York, (1990).

[45] N.G. Markley and M.E. Paul, Maximal measures and entropy for Zν sub-shifts of finite type. Classical mechanics and dynamical systems, Med-ford, Mass., (1979) 135-157, Lecture Notes in Pure and Appl. Math., 70, Dekker, New York, (1981).

[46] N.G. Markley and M.E. Paul, Matrix subshifts for Zν symbolic dynam-ics, Proc. London Math. Soc., (3) 43, (1981) 251-272.

[47] L. Onsager, Crystal statistics. I. A two-dimensional model with an order-disorder transition., Phys. Rev. 65, (1944) 117-149.

[48] C.N. Yang and C.P. Yang, One-dimensional chain of anisotropic spin-spin interactions. I. Proof of Bethe’s hypothesis for ground state in a finite system., Phys. Rev. 150, (1966a) 321-327.

[49] C.N. Yang and C.P. Yang, One-dimensional chain of anisotropic spin-spin interactions. II. Properties of the ground-state energy per lattice site for an infinite system., Phys. Rev. 150, (1966b) 327-339.

[50] C.N. Yang and C.P. Yang, One-dimensional chain of anisotropic spin-spin interactions. III. Applications., Phys. Rev. 151, (1966c) 258-264.

在文檔中 Pattern Generation Problems (頁 117-124)

相關文件