arXiv subjects
Publications and source records attributed to Song-Sun Lin.
Topological entropy or spatial entropy is a way to measure the complexity of shift spaces. This study investigates the relationships between the spatial entropy and the various periodic entropies which are computed by skew-coordinated systems $\gamma\in GL_2(\mathbb{Z})$ on two dimensional shifts of finite type.
This investigation completely classifies the spatial chaos problem in plane edge coloring (Wang tiles) with two symbols. For a set of Wang tiles $\mathcal{B}$, spatial chaos occurs when the spatial entropy $h(\mathcal{B})$ is positive. $\mathcal{B}$ is called a minimal cycle generator if $\mathcal{P}(\mathcal{B})\neq\emptyset$ and $\mathcal{P}(\mathcal{B}')=\emptyset$ whenever $\mathcal{B}'\subsetneqq \mathcal{B}$, where $\mathcal{P}(\mathcal{B})$ is the set of all periodic patterns on $\mathbb{Z}^{2}$ generated by $\mathcal{B}$. Given a set of Wang tiles $\mathcal{B}$, write $\mathcal{B}=C_{1}\cup C_{2} \cup\cdots \cup C_{k} \cup N$, where $C_{j}$, $1\leq j\leq k$, are minimal cycle generators and $\mathcal{B}$ contains no minimal cycle generator except those contained in $C_{1}\cup C_{2} \cup\cdots \cup C_{k}$. Then, the positivity of spatial entropy $h(\mathcal{B})$ is completely determined by $C_{1}\cup C_{2} \cup\cdots \cup C_{k}$. Furthermore, there are 39 equivalent classes of marginal positive-entropy (MPE) sets of Wang tiles and 18 equivalent classes of saturated zero-entropy (SZE) sets of Wang tiles. For a set of Wang tiles $\mathcal{B}$, $h(\mathcal{B})$ is positive if and only if $\mathcal{B}$ contains an MPE set, and $h(\mathcal{B})$ is zero if and only if $\mathcal{B}$ is a subset of an SZE set.
The commonly used spatial entropy $h_{r}(\mathcal{U})$ of the multi-dimensional shift space $\mathcal{U}$ is the limit of growth rate of admissible local patterns on finite rectangular sublattices which expands to whole space $\mathbb{Z}^{d}$, $d\geq 2$. This work studies spatial entropy $h_{\Omega}(\mathcal{U})$ of shift space $\mathcal{U}$ on general expanding system $\Omega=\{\Omega(n)\}_{n=1}^{\infty}$ where $\Omega(n)$ is increasing finite sublattices and expands to $\mathbb{Z}^{d}$. $\Omega$ is called genuinely $d$-dimensional if $\Omega(n)$ contains no lower-dimensional part whose size is comparable to that of its $d$-dimensional part. We show that $h_{r}(\mathcal{U})$ is the supremum of $h_{\Omega}(\mathcal{U})$ for all genuinely two-dimensional $\Omega$. Furthermore, when $\Omega$ is genuinely $d$-dimensional and satisfies certain conditions, then $h_{\Omega}(\mathcal{U})=h_{r}(\mathcal{U})$. On the contrary, when $\Omega(n)$ contains a lower-dimensional part, then $h_{r}(\mathcal{U})<h_{\Omega}(\mathcal{U})$ for some $\mathcal{U}$. Therefore, $h_{r}(\mathcal{U})$ is appropriate to be the $d$-dimensional spatial entropy.
This study investigates the natural or intrinsic measure of a symbolic dynamical system $\Sigma$. The measure $\mu([i_{1},i_{2},...,i_{n}])$ of a pattern $[i_{1},i_{2},...,i_{n}]$ in $\Sigma$ is an asymptotic ratio of $[i_{1},i_{2},...,i_{n}]$, which arises in all patterns of length $n$ within very long patterns, such that in a typical long pattern, the pattern $[i_{1},i_{2},...,i_{n}]$ appears with frequency $\mu([i_{1},i_{2},...,i_{n}])$. When $\Sigma=\Sigma(A)$ is a shift of finite type and $A$ is an irreducible $N\times N$ non-negative matrix, the measure $\mu$ is the Parry measure. $\mu$ is ergodic with maximum entropy. The result holds for sofic shift $\mathcal{G}=(G,\mathcal{L})$, which is irreducible. The result can be extended to $\Sigma(A)$, where $A$ is a countably infinite matrix that is irreducible, aperiodic and positive recurrent. By using the Krieger cover, the natural measure of a general shift space is studied in the way of a countably infinite state of sofic shift, including context free shift. The Perron-Frobenius Theorem for non-negative matrices plays an essential role in this study.
This investigation studies the decidability problem of plane edge coloring with three symbols. In the edge coloring (or Wang tiles) of a plane, unit squares with colored edges that have one of $p$ colors are arranged side by side such that the touching edges of the adjacent tiles have the same colors. Given a basic set $B$ of Wang tiles, the decision problem is to find an algorithm to determine whether or not $\Sigma(B)\neq\emptyset$, where $\Sigma(B)$ is the set of all global patterns on $\mathbb{Z}^{2}$ that can be constructed from the Wang tiles in $B$. When $p\geq 5$, the problem is known to be undecidable. When $p=2$, the problem is decidable. This study proves that when $p=3$, the problem is also decidable. $\mathcal{P}(B)$ is the set of all periodic patterns on $\mathbb{Z}^{2}$ that can be generated by the tiles in $B$. If $\mathcal{P}(B)\neq\emptyset$, then $B$ has a subset $B'$ of minimal cycle generators such that $\mathcal{P}(B')\neq\emptyset$ and $\mathcal{P}(B")=\emptyset$ for $B"\subsetneqq B'$. This study demonstrates that the set $\mathcal{C}(3)$ of all minimal cycle generators contains $787,605$ members that can be classified into $2,906$ equivalence classes. $\mathcal{N}(3)$ is the set of all maximal non-cycle generators: if $B\in \mathcal{N}(3)$, then $\mathcal{P}(B)=\emptyset$ and $\mathcal{P}(\tilde{B})\neq\emptyset$ for $\tilde{B}\supsetneqq B$. The problem is shown to be decidable by proving that $B\in \mathcal{N}(3)$ implies $\Sigma(B)=\emptyset$. Consequently, $\Sigma(B)\neq\emptyset$ if and only if $\mathcal{P}(B)\neq\emptyset$.
We study one-dimensional motions of polytropic gas governed by the compressible Euler equations. The problem on the half space under a constant gravity gives an equilibrium which has free boundary touching the vacuum and the linearized approximation at this equilibrium gives time periodic solutions. But it is not easy to justify the existence of long-time true solutions for which this time periodic solution is the first approximation. The situation is in contrast to the problem of free motions without gravity. The reason is that the usual iteration method for quasilinear hyperbolic problem cannot be used because of the loss of regularities which causes from the touch with the vacuum. Interestingly, the equation can be transformed to a nonlinear wave equation on a higher dimensional space, for which the space dimension, being larger than 4, is related to the adiabatic exponent of the original one-dimensional problem. We try to find a family of solutions expanded by a small parameter. Applying the Nash-Moser theory, we justify this expansion.The application of the Nash-Moser theory is necessary for the sake of conquest of the trouble with loss of regularities, and the justification of the applicability requires a very delicate analysis of the problem.
This study investigates a multiplicative integer system using a method that was developed for studying pattern generation problems. The entropy and the Minkowski dimensions of general multiplicative systems can thus be computed. A multi-dimensional decoupled system is investigated in three main steps. (I) Identify the admissible lattices of the system; (II) compute the density of copies of admissible lattices of the same length, and (III) compute the number of admissible patterns on the admissible lattices. A coupled system can be decoupled by removing the multiplicative relation set and then performing procedures similar to those applied to a decoupled system . The admissible lattices are chosen to be the\ maximum graphs of different degrees which are mutually independent. The entropy can be obtained after the remaining error term is shown to approach zero as the degree of the admissible lattice tends to infinity.
The degree of mixing is a fundamental property of a dynamical system. General multi-dimensional shifts cannot be systematically determined. This work introduces constructive and systematic methods for verifying the degree of mixing, from topological mixing to strong specification (or strong irreducibility) for two-dimensional shifts of finite type. First, transition matrices on infinite strips of width $n$ are introduced for all $n\geq 2$. To determine the primitivity of the transition matrices, connecting operators are introduced to reduce the order of high-order transition matrices to yield lower-order transition matrices. Two sufficient conditions for primitivity are provided; they are invariant diagonal cycles and primitive commutative cycles of connecting operators. After primitivity is established, the corner-extendability and crisscross-extendability are used to demonstrate topological mixing. In addition, the hole-filling condition yields the strong specification. All mentioned conditions can be verified to apply in a finite number of steps.