アブストラクト事後公開 — 2018年度年会(於:東京大学)
応用数学分科会
特別講演 距離集合における分類問題と極値問題 篠原雅史 (滋賀大教育) We consider good configurations from the point of view of the theory of distance sets. A subset $X$ of a Euclidean space is called a $k$distance set if there exists exactly $k$ values of distances between two distinct points in $X$. The study of distance sets was initiated by Erdös (1946). One of the major problems in the theory of $k$distance sets in the $d$dimensional Euclidean space is to determine the largest possible cardinality $g_d(k)$ of $k$distance sets and classify the distance sets $X$ satisfying $X=g_d(k)$. Such $k$distance sets are said to be optimal. Furthermore, we are also interested in characterization of $k$distance sets with large points relative to $k$. In this talk, we introduce some results for optimal $k$distance sets and extremal problems on distance sets. msjmeeting2018mar09i001 

特別講演 構造保存型差分解法のエネルギー法 吉川周二 (大分大理工) We introduce the energy method for structurepreserving finite difference schemes which inherit the physical structures such as energy conservation or dissipative laws. Another aim is to give some useful properties for difference quotient which is compatible with the structurepreserving finite difference schemes. The method and properties enable us not only to take the problem with more general nonlinearity but also to improve proofs of error estimate between the numerical and exact solutions. In this talk, after explaining our procedure by using a simple example, several our recent results are introduced. msjmeeting2018mar09i002 

特別講演 半群理論を利用する発展方程式に対する解の精度保証付き数値計算 高安亮紀 (筑波大システム情報) In this talk, we provide a methodology of verified computing for solutions to evolution equations (nonlinear heat equations, 1dimensional advection equations with variable coefficients, and the complex Ginzburg–Landau equations). Our methodology is based on semigroup theory, which is widely used in analytical studies and originated from pioneering works by Hille and Yosida. The main contribution of this study is to combine a “classical analysis” with “computerassisted methods” to provide a numerical method of enclosing a solution for evolution equations. The combination of quantitative estimates arising from verified numerical computations and qualitative results obtained by classical analysis is expected to open the access to many unsolved problems by purely analytical means. msjmeeting2018mar09i003 

1. 
サイクリックになる二部グラフのラベル付けとRAIDへの応用 足立智子 (東邦大理) A cluttered ordering is a kind of cyclic orderings, and can be used to minimize the number of disk operations in RAID system. Mueller et al. (2005) decomposed the complete bipartite graph into isomorphic copies of the special bipartite graph $H(h; t)$, where $h$ and $t$ are positive integers. The special bipartite graph $H(h; t)$ has each $h(t+1)$ vertices as upper vertex set and lower vertex set. In this talk, we define the special bipartite graph $H(h, k; t)$, where $h$, $k(h \neq k)$ and $t$ are positive integers. The special bipartite graph $H(h, k; t)$ has $h(t + 1)$ vertices as upper vertex set and $k(t + 1)$ vertices as lower vertex set. We present some edge labeling of the infinite families of $H(1, 2; t)$. 

2. 
Vietoris–Rips realization of indecomposable persistence modules of arbitrarily large dimension M. Buchet (東北大AIMR／TU Graz)・E. G. Escolar (東北大AIMR) In this work we construct an infinite family, parametrized by dimension order $d$, of nonisomorphic indecomposable persistence modules over the commutative ladder of length 5. We provide a family of bifiltrations of topological spaces whose $H_1$ persistent homologies is the infinite family of persistence modules. Moreover, we provide Vietoris–Rips constructions of the family of bifiltrations. Our construction provides an elementary proof of the fact that the commutative ladders with length greater than or equal to 5 is representationinfinite. Furthermore, we aim by this example to illustrate that indecomposable persistence modules of high dimension as a representation may encode some interesting and easy to visualize phenomenon. 

3. 
Application of the theory of quasiorthogonal integrals to graph theory 瀬戸道生 (防衛大)・須田 庄 (愛知教育大) In this talk, we give two inequalities concerning increasing families of finite graphs. Those inequalities are derived from the theory of quasiorthogonal integrals developed by de Branges–Rovnyak and Vasyunin–Nikol’skii. 

4. 
Remark on two typical properties of random graphs 佐竹翔平 (神戸大システム情報) There are some well known properties which random graphs satisfy with probability $1$. Here we deal with two such typical properties of random graphs, namely, the $n$existentially closed ($n$e.c.) property and the pseudorandom property. Cameron and Stark remarked that the pseudorandom property does not necessarily imply the $n$e.c. property for large $n$. However there seems no results about the inverse relationship. In this talk, we give the first construction of $n$e.c. graphs without the pseudorandom property, which means that the $n$e.c. property also does not necessarily imply the pseudorandom property for every $n$. 

5. 
ランダム複体過程における生存時間和の漸近評価 金澤 秀 (東北大理)・日野正訓 (京大理) We study the homological properties of random simplicial complexes, which have received a lot of attention over the past several years. In particular, we obtain asymptotic behavior of lifetime sums of persistent homology for a class of increasing random simplicial complexes, which is a higherdimensional counterpart of Frieze’s zeta function theorem for the Erdős–Rényi graph process. Main results include solutions to the questions on the Linial–Meshulam complex process and the clique complex process that were posed in the preceding study by Hiraoka and Shirai. One of the key ingredients of the arguments is a new upper bound of Betti numbers of general simplicial complexes in terms of the number of small eigenvalues of Laplacians on links, which is regarded as a quantitative version of the cohomology vanishing theorem. 

6. 
軸平行な直方体内にあるChebyshev–Frolov格子点の列挙アルゴリズム 鈴木航介 (広島大理)・芳木武仁 (京大情報) For a positive integer $d$, the $d$dimensional Chebyshev–Frolov lattice is the $\mathbb{Z}$lattice in $\mathbb{R}^d$ generated by the Vandermonde matrix associated to the roots of the $d$dimensional Chebyshev polynomial. It is important to enumerate the points from the Chebyshev–Frolov lattices in axisparallel boxes when $d = 2^n$ for a nonnegative integer $n$, since the points are used for the nodes of Frolov’s cubature formula, which achieves the optimal rate of convergence for many spaces of functions with bounded mixed derivatives and compact support. The existing enumeration algorithm for such points by Kacwin, Oettershagen and Ullrich is efficient up to dimension $d=16$. In this paper we suggest a new enumeration algorithm of such points for $d=2^n$, efficient up to $d=32$. 

7. 
位相空間の圏と同型な関係T代数について 阿川真士 (九大数理)・溝口佳寛 (九大IMI) A continuous lattice is a semantic domain of a computation such as a lambda calculous. Our motivation comes from a development of a formal theory of semantic domains. We first talk about a theory of relational Talgebra, an extension of a Talgebra. Michael Barr proved the category of relational Talgebra defined by the ultrafilter monad is isomorphic to the category of topological spaces. We review their theory and reformulate it using our simple flamework of a relational calculus. 

8. 
The Smith normal form of skewsymmetric Doptimal designs of order $n\equiv2\pmod{4}$ 須田 庄 (愛知教育大)・G. Greaves (Nanyang Tech. Univ.) We show that the Smith normal form of a skewsymmetric Doptimal design of order $n\equiv 2\pmod{4}$ is determined by its order. We apply our result to show that certain Doptimal designs of order $n\equiv 2\pmod{4}$ are not equivalent to any skewsymmetric Doptimal design. 

9. 
Maximizing the order of a bipartite regular graph for given valency and second eigenvalue 野﨑 寛 (愛知教育大) Let $v(k,\lambda)$ be the maximum order of connected bipartite $k$regular graphs whose secondlargest eigenvalues are at most $\lambda$. We show an upper bound for $v(k,\lambda)$, which is based on the linear programming bound. If a graph attains the bound, then it is a distanceregular graph that satisfies $g\geq 2d2$, where $g$ is the girth and $d$ is the diameter of the graph. There are infinitely many bipartite distanceregular graphs that satisfy $g\geq 2d2$. We can prove the nonexistence of bipartite distanceregular graphs with $g\geq 2d2$ for $d>26$ by the manner of Fuglister (1987). This is a joint work with Sebastian Cioabă and Jack Koolen. 

10. 
Kernels by properly colored paths in arccolored digraphs Yandong Bai (Northwestern Polytechnical Univ.)・藤田慎也 (横浜市大国際総合)・Shenggui Zhang (Northwestern Polytechnical Univ.) A subdigraph $H$ of an arccolored digraph $D$ is properly colored if any two consecutive arcs of $H$ receive distinct colors. A kernel by properly colored paths of an arccolored digraph $D$ is a set $S$ of vertices of $D$ such that (i) no two vertices of $S$ are connected by a properly colored directed path in $D$, and (ii) every vertex outside $S$ can reach $S$ by a properly colored directed path in $D$. We conjecture that every arccolored digraph with all cycles properly colored has such a kernel and verify the conjecture for unicyclic digraphs, semicomplete digraphs and bipartite tournaments, respectively. 

11. 
Distance matching extension in cubic bipartite graphs 藤沢 潤 (慶大商)・R. E. L. Aldred (Univ. of Otago)・斎藤 明 (日大文理) A matching $M$ of a graph $G$ is said to be extendable if $M$ is a subset of a perfect matching of $G$, and $M$ is said to be distance $d$ matching if the edges of $M$ lie pairwise distance at least $d$. If every distance $d$ matching of $G$ is extendable, then we say that $G$ is distance $d$ matchable. In this talk we introduce the following results: 1) Let $G$ be a $3$connected cubic bipartite graph. If there exist two cycles $C_1$, $C_2$ of length at most $d$ such that $E(C_1) \cap E(C_2) = \{e\}$ for every $e \in E(G)$, then $G$ is distance $d$ matchable. 2) Let $G$ be a $3$connected cubic bipartite planar graph. If there exist two cycles $C_1$, $C_2$ of length at most $6$ such that $e \in E(C_i)$ ($i=1,2$) for every $e \in E(G)$, then $G$ is distance $6$ matchable. 

12. 
トーラス上の三角形分割の$n$triad coloring 大野由美子 (横浜国大環境情報) A triangulation on a closed surface is a graph embedded on the surface each of whose face is triangular. Let $G$ be a triangulation on a closed surface and $n \ge 3$ be a natural number. A coloring $c : V(G) \to \mathbb{Z}_n$ is called an $n$triad coloring if $\{c(u), c(v), c(w)\}$ belongs to $\{\{i, i+1, i+2\} \mid i \in \mathbb{Z}_n\}$ for any face $uvw$ of $G$. We would like to determine the set of numbers $n$ such that $G$ has $n$triad colorings. The set can be determined completely by the chromatic number of $G$ if $G$ is embedded on the sphere or the projective plane. In this talk, we shall focus on $G$ which is embedded on the torus and investigate the above set of $G$. 

13. 
3dynamic coloring for triangulations on surfaces 朝山芳弘 (横浜国大環境情報)・川崎雄貴 (横浜国大環境情報)・SeogJin Kim (Konkuk Univ.)・中本敦浩 (横浜国大教育人間)・小関健太 (横浜国大環境情報) An rdynamic kcoloring of a graph $G$ is a proper $k$coloring such that any vertex $v$ has at least min$\{r,\deg_G(v)\}$ distinct colors in $N_G(v)$. The rdynamic chromatic number $\chi_r^d(G)$ of a graph $G$ is the least $k$ such that there exists an $r$dynamic $k$coloring of $G$. Loeb and et al, proved that $\chi^d_3\leq10$ if $G$ is a planar graph, however this result is not considered not to be sharp. Thus finding an optimal upper bound on $\chi^d_3(G)$ for a planar graph $G$ is a natural interesting problem. We will show some upper bounds on $\chi^d_3(G)$ for triangulations on the plane, the projective plane and the torus. 

14. 
球面以外の閉曲面に埋め込まれた3連結3正則平面的グラフ 永並健吾 (横浜国大環境情報) It is wellknown that every 3connected planar graph is uniquely embeddable on the sphere but it is not uniquely embeddable on any surface other than the sphere. We shall focus on a 3connected 3regular planar graph and classify structures of its embeddings on the torus, the projective plane and the Klein bottle. 

15. 
射影平面上の正則グラフの符号 小関健太 (横浜国大環境情報)・阿部敏生 (横浜国大環境情報) Let $G$ be a $k$regular $k$edge colorable graph. Moreover, let $\varphi$ be a $k$edge coloring of $G$. Let $v$ be a vertex of $V(G)$ and $E(v)$ denotes the set of all edges incident with $v$. Let $\rho_v$ be a bijective map $E(v) \to \{1,2,...,k\}$ and we call the set $\rho = \{\rho_v: v \in V(G)\}$ basis. Now we obtain the correspondence $\pi_v = \varphi \circ \rho_v^{1}$ for every $v$. The signature of $\varphi$ is defined as $\prod_{v \in V(G)}$ sign$(\pi_v)$. In this talk, we obtain the signatures of $k$edgecolorings in $k$regular graphs on the projective plane for a certain basis $\rho$. Moreover, this has an application to the list coloring conjecture for certain graph class by using former result. 

16. 
偶三角形分割における辺数の大きな二部的全域部分グラフ 野口健太 (東京電機大情報) We consider a spanning bipartite subgraph of an even (i.e. Eulerian) triangulation $G$ on a surface. If $G$ has a spanning bipartite quadrangulation, then its size is twothirds of $E(G)$. In this talk, we discuss the condition of $G$ having a spanning bipartite quadrangulation $Q$. We also discuss the maximum size of a spanning bipartite subgraph of $G$ when $G$ does not have $Q$. 

17. 
例外的3染色的三角形分割について 鈴木有祐 (新潟大理)・村井 聡 (阪大情報) In our recent research, we proved that two balanced (or 3colorable) triangulations of a closed surface are not necessary connected by a sequence of balanced stellar subdivisions and welds. This answers a question posed by Izmestiev, Klee and Novik. In this talk, we especially discuss two local operations called a pentagon contraction and a pentagon splitting, which are also defined for balanced triangulations of closed surfaces. We show that most two balanced triangulations of a closed surface are transformed into each other by a sequence of the above two operations. Furthermore, we introduce such exceptional balanced triangulations of closed surfaces with low genera. 

18. 
A weighted generalized Bartholdi zeta function of a digraph 佐藤 巖 (小山工高専)・三橋秀生 (法政大理工)・森田英章 (室蘭工大工) We define a weighted generalized Bartholdi zeta function and a weighted generalized Bartholdi $L$function of a digraph, and present their determinant expressions. Furthermore, we give express the weighted generalized Bartholdi zeta function of a group covering of a digraph by a product of its weighted generalized Bartholdi $L$functions. 

19. 
空間的に非一様な 1 次元 2 状態の量子ウォーク 井手勇介 (神奈川大工)・今野紀雄 (横浜国大工)・中山大地 (横浜国大工) We consider the twostate spaceinhomogeneous coined quantum walk (QW) in one dimension. We obtain the uniform measure as the stationary measure by solving the eigenvalue problem. This approach is based on the method giving by Kawai, Komatsu and Konno. (2017). 

20. 
2次元トーラス上の量子ウォーク 小松 尭 (横浜国大理工)・今野紀雄 (横浜国大工)・成松明廣 (横浜国大工) We analyze two types of the Quantum walk on the 2dimensional torus, the Grover walk and the Fourier walk, and obtain provability amplitude and the search algorithm using the Fourier walk. 

21. 
2次元スプリットステップ量子ウォークの固有値解析 船川大樹 (北大理)・布田 徹 (北大理)・笹山智司 (北大理)・鈴木章斗 (信州大工) We consider the 2dimensional 4states quantum walk. This quantum walk is an extension of the 1dimensional splitstep quantum walk. By the earlier study, it is known that if a particular function $f$ has zero points, then time evolution operator $U$ has eigenvalues. In such a case, localization occurs. In this talk, we introduce the necessary and sufficient condition of $f$ has zero points. This research is a joint work with T. Fuda, S. Sasayama and A. Suzuki. 

22. 
1次元系上のものぐさフーリエウォーク 寺田知幸 (金沢工大)・小松 尭 (横浜国大理工)・今野紀雄 (横浜国大工)・齋藤 渓 (横浜国大工) We consider the property of lazy Fourier walk in one dimension. For example, stationary measure, timeaveraged limit measure, and periodicity. 

23. 
1次元量子ウォークのユニタリ同値類 大野博道 (信州大工) This study investigates the unitary equivalence classes of onedimensional quantum walks. We determine the unitary equivalence classes of onedimensional quantum walks, twophase quantum walks with one defect, complete twophase quantum walks, onedimensional quantum walks with one defect and translationinvariant quantum walks. 

24. 
高次元整数格子上の量子ウォークの定常測度 小松 尭 (横浜国大理工)・今野紀雄 (横浜国大工) Stationary measures of quantum walks on the onedimensional integer lattice are well studied. However, the stationary measure for the higher dimensional case has not been clarified. In this talk, we give the stationary amplitude for quantum walks on the higherdimensional integer lattice with a finite support by solving the corresponding eigenvalue problem. As a corollary, we can obtain the stationary measures of the Grover walks. 

25. 
空間非一様な量子ウォークの定常測度 小松 尭 (横浜国大理工)・川合 光 (横浜国大工)・今野紀雄 (横浜国大工) We consider the twostate spaceinhomogeneous coined quantum walk in one dimension. For a general setting, we obtain the stationary measures of the quantum walks by solving the eigenvalue problem. As a corollary, stationary measures of the multidefect model and spacehomogeneous quantum walks are derived. 

26. 
グラフのジョイントとグローヴァーウォークの周期性 吉江佑介 (東北大情報) The Grover walk is a kind of quantum walks on graphs, and it is applied to various study fields. The Grover walk is determined by a unitary time evolution operator given by the underlying graph. So we can say that the Grover walk is induced by the graph. We focus on characterizations of graphs inducing periodic Grover walks, that is, there exists a integer $k$ such that $k$th iteration of the time evolution operator becomes identity operator. In previous walk, such graphs have been found, e.g. cycle graphs, path graphs, complete bipartite graphs. In this talk, we construct new graphs with these graphs, and consider the periodicity of the Grover walk on these graphs. 

27. 
Persistence weighted Gaussian kernel によるブートストラップ法 草野元紀 (東北大理) In topological data analysis, the topological structures in data are transformed into a persistence diagram, and its statistical method is proposed by the Persistence weighted Gaussian kernel (PWGK). Here, let us consider the expectation of persistence diagrams by the PWGK. While it is difficult to calculate the expectation in general, the confidence set which contains the true expectation can be constructed by the bootstrap method. In this talk, we will discuss the bootstrap method by the PWGK and its asymptotic consistency. 

28. 
対応のパーシステントホモロジー: 箙の表現論の視点から 竹内博志 (東北大理)・平岡裕章 (東北大AIMR) The theory of homology induced maps of correspondences proposed by Shaun Harker et al. is a powerful tool which allows the retrieval of underlying homological information from sampling data with noise or defects. In this study, we redefine induced maps of correspondences within the framework of quiver representations, and provide more concise proofs of the main theorems in the original paper. With this point of view, we easily extend these ideas to filtration analysis, which provides a new method for analyzing dynamical systems. 

29. 
Volume optimal cycles for persistent homology 大林一平 (東北大AIMR) Persistent homology (PH) is a significant tool for topological data analysis, which analyzes shape of data efficiently and quantitatively. A persistence diagram (PD) is a visualization tool of PH, which is a multiset on $\mathbb{R} \times (\mathbb{R}\cup\infty)$. Each point on a PD (called a birthdeath pair) corresponds a homological structure such as a ring, cavity, etc. appearing in the data. For a practical application of PH, we want to identify such a homological structure for a selected birthdeath pair. The idea of a volume optimal cycle and the computation algorithm are proposed in this talk. 

30. 
The case studies of using computational homology in medical science 寺本 敬 (旭川医科大医)・瀧山晃弘 (北海道文教大人間科学) We present a methodlogy using computation homology for quantitatve measurements in medical science, in collaboration with diagnostic doctors. Our talk will consits of two folds: cubical homology index for bone morphometry in threedimension, and immunohistochemical scoring based on persistent homology. 

31. 
The Conley index over base 岡 宏枝 (龍谷大理工)・神山蒼太 (龍谷大理工) The Conley index over a base [Mrozek Reineck, Srzednicki 1997] is a generalization of the Conley index for flows [Conley 1972]. In the case when the base is the circle, it is naturally related to the discrete type of Conley index of a Poincare section of the flow. We will consider the homology version of the Conley index over a base, and discuss the relation between them. 

32. 
拡張複素ニュートン法の根の収束と分布について 堀口俊二 We extend the complex Newton’s method. We give the followings for the extended complex Newton’s method. Relationship between extended complex Newton’s method and Riemann surface. Conditional expression of initial values for convergence of extended complex Newton’s method. Speeds of convergences of extended complex Newton’s method. The distributions of roots of extended complex Newton’s Method. 

33. 
微分方程式の整数型解法における余剰解の整数数列の「解読」 坂口文則 (福井大工) An integertype algorithm for solving ODEs was proposed by the author and M. Hayashi. This algorithm is based on the expansion of solution functions by rationalfunctiontype basis functions, and it is based on the ‘exact’ kernel vectors of nonsquare matrices. In this algorithm, we can read and ‘decipher’ integer coefficient sequences directly, and hence we can analyze behavior of numerical solutions exactly and puremathematically. In this study, the author gives further numerical examples where it is clearly shown that many hyperfunction components are contained in extra solutions mixed in numerical solutions obtained by this algorithm, by a direct ‘decipherment’ of integer coefficient sequences contained in numerical results by this algorithm, than the examples already presented. 

34. 
Second order unbiased simulation method for reflected stochastic differential equations 湯浅智意 (立命館大理工)・秋山樹生 (立命館大理工)・A. KohatsuHiga (立命館大理工) In this talk, we consider unbiased simulation methods functionals of solutions of onedimensional reflected stochastic differential equations. Alfonsi–Hayashi–Kohatsu propose an unbiased simulation method for the present problem based on the parametrix method. However, the variance of this method is not finite in general unless one uses an importance sampling method. We propose a different way of obtaining what we call a second order parametrix method which leads to an alternative unbiased random variable with finite moments. We call this method “Second order unbiased simulation method for reflected stochastic differential equations”. 

35. 
値域が共役空間となる楕円型作用素に対する精度保証付き可逆性検証 木下武彦・渡部善隆 (九大情報基盤研究開発センター)・中尾充宏 (早大理工) This talk presents an computerassisted procedure for verifying the invertibility of secondorder linear elliptic operators and for computing a bound on the norm of its inverse. This approach is an improvement of a theorem (Nakao, et al. 2015, Jpn. J. Ind. Appl. Math. 32, 19–32) that uses projection and constructive a priori error estimates. Several examples which confirm the actual effectiveness of the procedure are reported on. 

36. 
Proudman–Johnson方程式のunimodalな定常解に対する精度保証付き数値計算 宮路智行 (明大MIMS)・岡本 久 (学習院大理) We study the existence of unimodal stationary solutions to the Proudman–Johnson equation with an external force. In particular, we are interested in the case of a high Reynolds number. In order to prove the existence and unimodality of a solution, we resort to interval arithmetic. We formulate the stationary problem for the Proudman–Johnson equation as a system of first order ordinary differential equations, and we apply the shooting method and the interval Newton method for proving the existence of a solution. As the shooting method is numerically unstable, we encounter some difficulties especially when the Reynolds number is high. For solving this problem, we apply the multiple shooting method and the multipleprecision floatingpoint arithmetic. 

37. 
放物型問題のNitsche’s methodに対するinfsup条件及び誤差評価について 上田祐暉 (東大数理)・齊藤宣一 (東大数理) We study the application of the Nitsche’s method to the parabolic problems. Under some assumptions, the parabolic initialboundary value problem has a unique weak solution. The problem is discretized in space by the Galerkin method and the Dirichlet boundary condition is enforced weakly by the Nitsche’s method. It is well known that the bilinear form satisfies the Galerkin orthogonality. In this presentation, we will prove that the bilinear form also satisfies the infsup condition. This condition implies that the resulting semidiscretized problem has a unique solution. Moreover, the error estimate follows directly from the infsup condition and the Galerkin orthogonality. 

38. 
重調和問題に対するある混合型非適合有限要素法への内部ペナルティ法の適用 小山大介 (電通大) We introduce a numerical method for biharmonic problems which is obtained by applying the interior penalty method to a mixed nonconforming finite element method which is called the Hermann–Johnson (HJ) method. We show that a priori error estimates of our method can be the same as those of the HJ method by appropriately choosing a penalty parameter in our method. 

39. 
3次元領域におけるStokes方程式の有限要素解の事前誤差評価 劉 雪峰 (新潟大自然)・中尾充宏 (早大理工／九大名誉教授)・遊 春光 (China Acad. of Eng. Phys.)・大石進一 (早大理工) For the Stokes equation defined in 3D domain with a general shape, the Scott–Vogelius finite elements are used to obtain strictly divergencefree approximation solution. Then, by using the hypercircle equation method, a quantitative a priori error estimation is obtained for the FEM solution. Such an a priori error estimation can be used in solution existence verification of nolinear Navier–Stokes equation defined in 3D domain with general shapes. The convergence rate is confirmed by numerical results. 

40. 
不連続Galerkin時間離散化法の変分法的な解析 齊藤宣一 (東大数理) The discontinuous Galerkin (DG) timestepping method applied to abstract evolution equation of parabolic type is studied using a variational approach. We establish the infsup condition for the DG bilinear form. Then, the optimal order error estimates under appropriate regularity assumption on the solution are derived as direct applications of the standard interpolation error estimates. Our method of analysis is new. It differs from previous works on the DG timestepping method by which the method is formulated as the onestep method. 

41. 
DG timestepping 法に対する解析半群論的なアプローチ 剱持智哉 (東大数理) The discontinuous Galerkin timestepping method (DG timestepping method) is a timediscretization method based on the discontinuous Galerkin finite element method. In contrast to onestep methods, the approximated solution is welldefined at each time in the DG timestepping method. Therefore, it gives an efficient numerical algorithm with spacetime methods for moving boundary problems such as fluid structure interaction. However, there are few studies on theoretical analysis for the behavior of approximated solutions at each time. In this talk, we address the DG timestepping method for parabolic problems in the framework of analytic semigroup theory. We present optimal order error estimates for the homogeneous heat equation. The key point is rigorous estimates for rational functions that express the approximated solutions. 

42. 
熱方程式の初期値問題(横向きを含む)に対する不安定差分スキームの数値計算例 東森信就 (京大CPIER) We previously reported examples of convergence of unstable finite difference schemes applied to quasilinear partial differential equations of the normal form with analytic initial data. This time we report that convergence also holds for initial value problems for the heat equation with initial data in a certain class of analytic functions and also for the sideways heat equation with initial data in a Gevrey class. 

43. 
De Sitter時空における半線形Klein–Gordon方程式の数値安定性について 土屋拓也 (早大理工)・中村 誠 (山形大理) We investigate the numerical stability for semilinear Klein–Gordon equations in de Sitter spacetime. We show the differences of the numerical stability between some positive nonlinear terms and the negative one. In addition, the numerical stability in the four dimension and more is shown. 

44. 
変分原理に基づくエネルギー保存数値解法のLie群上への拡張 石川歩惟 (神戸大システム情報)・谷口隆晴 (神戸大システム情報) The energypreserving method based on the variational principle and the discrete gradient method are methods for designing a scheme for the Hamilton equations that preserves the energy conservation law exactly. In the discrete gradient method, the various extensions, e.g. the extension to the equations on a manifold have been proposed. Although the method based on the variational principle has an advantage in the view of the computational cost, this method has not been applied to the equations on the manifold yet. In this talk, we extend the energypreserving method based on the variational principle to equations on the Lie group and show a numerical test for the heavy top problem. 

45. 
Nonlinear and linear DVDM scheme for the conservative nonlocal Allen–Cahn equation 奥村真善美 (阪大情報) We propose nonlinear and linear finite difference scheme for the conservative nonlocal Allen–Cahn equation. Both proposed schemes inherit characteristic properties, the conservation of mass and the decrease of the global energy from the equation. We show that the schemes are stable in the sense that the numerical solution is bounded concerning maxnorm, and have a unique solution. Since the nonlinear scheme is the system of equations concerning the new time step, it takes time to compute. Numerical examples demonstrate the effectiveness of the proposed scheme and that the computational time of the linear scheme is shorter than one of the nonlinear scheme. In this talk, we mainly introduce the linear scheme. 

47. 
自由境界をもつある感染症モデルについて 江夏洋一 (東京理大理)・石渡恵美子 (東京理大理)・牛島健夫 (東京理大理工) Free boundary problems are recently used to model phenomena of biological invasion for species such as migration into a new habitat (e.g., Du & Lin (2010) and references therein). These ideas are also applied to epidemic models. In this talk, we extend the result in Kaellen (2017) to the simple diffusive epidemic model with free boundary, namely we prove the existence of a semi wave solution. We numerically observe the semi wave and the front motion of this model with free boundary. 

48. 
対流の効果を含む界面活性剤系自走粒子の自律運動に関する数理モデル 岡本 守 (北大理)・長山雅晴 (北大電子研) Many researchers have studied the selfdriven particles. In one example, there is camphor atop water channel. It is now said that the motion of camphor is caused by differences in surface tension. The gradient of surface tension is induced by a camphor molecular layer development atop the surface. Mathematical models for the camphor motion have been constructed used the above mechanisms, and the models reproduce many characteristic motion. Although Marangoni Convection seems to influence the selfmotion of the camphor, there are only a few reports discussing mathematical models that include convection explicitly. We have constructed a mathematical model for the selfmotion of camphor including influence of convection, and now are calculating some cases to compare with some experiments. 

49. 
表皮構造の数理モデルにおける顆粒層の安定化とタイトジャンクション形成への取り組み 後藤田剛 (北大電子研)・上坂正晃 (北大電子研)・安ケ平裕介 (北大理)・小林康明 (お茶の水女大理)・北畑裕之 (千葉大理)・傳田光洋 ((株)資生堂)・長山雅晴 (北大電子研) We propose a mechanism of stable formation of the granular layer in the epidermis. In our mathematical model of the epidermis, we assume that a stimulant, which promotes the differentiation process of epidermal cells, is released when cells undergo cornification. We demonstrate that our model forms the granular layer and confirm that its layer structure is maintained stably by using some cost functions. We are also working on formation of tight junctions, which exit in the granular layer and play an important role in skin barrier function, and we introduce our trial models for them. 

50. 
基底膜に覆われた細胞群の回転運動を表現するための数理モデル 須志田隆道 (北大電子研)・森 瞳 (北大理)・石田すみれ (北大生命)・古澤和也 (北大先端生命)・芳賀 永 (北大先端生命)・秋山正和 (北大電子研) Recently, it has been attended the relation between the elongation phenomena and rotational migration of cell group since they are observed in threedimensional morphogenesis such as fruiting body formation of Dictyostelium Discoideum and somite formation of zebra fish. In particular, although it is known that the somite is covered by basement membrane, it is not clear the reason why rotational migration occurs. In order to understand cellular mechanism for rotational migration of cell groups covered by basement membrane, we propose a mathematical model which consists of a selfpropelled particle model representing cellular migration and a phasefield model representing basement membrane. Moreover, we will show phase diagram of parameters for migration modes and give a theoretical suggestion for biological experiments. 

51. 
混合微分を含む発展方程式の再定式化について 佐藤 峻 (東大情報理工) In this talk, we deal with the initial value problem of evolutionary differential equations with a mixed derivative on the periodic domain. Here, “mixed derivative” indicates the case where a spatial differential operator is operating on the time derivative, obscuring the vector field describing the flow. Therefore, some reformulation to reveal it is the first step of PDEtheoretical and numerical studies. However, it is nontrivial because the spatial differential operator is not invertible and cannot be easily eliminated. Though this issue was already settled for linear cases, general theory has been undeveloped. In this talk, we propose a novel procedure for wider class of equations. Moreover, as an application, we establish the global wellposedness of the sineGordon equation in characteristic coordinates. 

52. 
対称曲面上のGreen関数 清水雄貴 (京大理) Uniform flow is one of fundamental steady solutions of Euler equation on a plane. A generalization of the flow on curved surfaces is a Killing vector field, which is also a steady solutions of Euler equation on Riemannian manifolds. In this talk, we introduce how surfaces has a nonormal regular Killing vector field and as its application, construction of an analytic formula of Green’s function on the surface. 

53. 
非粘性保存量の乱流カスケードを実現する1次元流体方程式について 坂上貴之 (京大理)・松本 剛 (京大理) We propose a onedimensional hydrodynamic partial differential equation. This model is based on a Constantin–Lax–Majda–De Gregorio model generalized by Okamoto, Sakajo and Wunsch. The equation admits an inviscid invariant quantity. In the presence of the viscosity and a largescale random forcing, the solution gives rise to a turbulent state with cascade of the inviscid invariant. We will give how those phenomenon is understood from the view point dynamical system. 

54. 
Kasper翼まわりの定常点渦の安定化を実現する線型フィードバック制御 坂上貴之 (京大理)・R. Nelson (Imperial Coll. London)・B. Protas (McMaster Univ.) Based on the LinearQuasiGaussian compensator, we design a linear feedback system stabilizing point vortex equilibria near an aerodynamic wing with two auxiliary flaps known as a Kasper Wing in the presence of a uniform flow. This is modeled by a twodimensional incompressible an inviscid flow. The actuation mechanism is blowing and suction localized on the main plate represented as a sink source singularity, whereas we measure pressure across the plate as system output. We show that the linearised system around these equilibria are both controllable and observable for almost all actuator and sensor locations. Numerical computations illustrate that Kasper Wing configurations are in general not only more controllable than their single plate counterparts, but also acquire larger basins of attraction owing to the feedback control. 

55. 
一般J積分による固有値の形状感度解析 大塚厚二 (広島国際学院大情報文化) Generalized Jintegral is the tool which is effective to study the shape optimization of singular points (containing boundary) with respect to various cost functions, energy, mean compliance, least square errors, in boundary value problems for partial differential equations. I will talk an application of Generalized Jintegral method to shape sensitivity of eigenvalue problems. 

56. 
自発的なパラメータ制御による集団振動子の同期現象 上田肇一 (富山大理) We study synchronization of two interacting populations of oscillators. We assume that the sign and the strength of the interactions are taken as system variables, and that the variables are determined by a function of the oscillator variables. Under an appropriate choice of the function, we can observe spontaneous intra and intergroup phase synchronization. 

57. 
振動テールパルスと非一様媒質の相互作用を支配する大域的分岐構造 渡辺 毅 (東大工)・Zhijun Gao (東北大AIMR)・西浦廉政 (東北大AIMR) Threecomponent FitzHugh–Nagumo equation is investigated. This equation has a parameter region in which traveling pulse with oscillatory tail appears. When such moving pulse interacts with heterogeneity of the media, the interaction between tail and heterogeneity is important. We consider a bumptype heterogeneity and investigate the asymptotic behavior of the pulse motion when a pulse collides with the bump. When the width of the bump is fixed and the height of that is changed, three different asymptotic behaviors arise sequentially: Oscillatory pinning (OSC), stationary pinning (STA), and rebound (REB). In these asymptotic states, it is conjectured that all asymptotic states are contained in the set of bifurcation branches which is generated from the trivial branch. 

58. 
円環水路上の2個の樟脳ろ紙がみせる運動の分岐解析 西 慧 (京都産大理)・長山雅晴 (北大電子研)・中田 聡 (広島大理) The dynamics of two selfpropelled camphor disks on an annular field is explored both numerically and analytically. In our previous study, it was already found by the direct numerical simulation of a model equation that the two camphor disks exhibited a variety of behavior, and underwent transition between the behavior as the length of the annular field was varied. In order to analytically elucidate the mechanism for the transition in behavior, we reduced the model equation which consisted of two ODEs and one PDE into three ODEs for the motion of the two camphor disks. In this talk, the bifurcation structure that causes the transition will be revealed, based on the reduced ODEs. 

59. 
2自由度ハミルトン系における横断的なヘテロクリニック軌道の存在と非可積分性 山中祥五 (京大情報)・矢ヶ崎一幸 (京大情報) We consider a class of twodegreeoffreedom Hamiltonian systems with saddlecenters connected by heteroclinic orbits. We show that if the sufficient conditions for realmeromorphic nonintegrability hold, then the stable and unstable manifolds of the periodic orbits intersect transversely, are quadratically tangent or do not intersect in general, and they do not intersect when the Hessian matrix of the Hamiltonian has a different number of positive eigenvalues at the associated saddlecenters. Our theory is illustrated for a system with quartic singlewell potential. 

60. 
導函数座標埋め込みによって得られる曲面の性質と函数の時間発展の予測可能性について 中野直人 (京大国際高等教育院／JSTさきがけ) Here, we consider the derivative embedding of a scalar function. Since the embedded surface can hold some characteristics of the original timeseries, one can derive some predictability result from characteristics of the embedded surface. 

61. 
1次元非線形波動方程式の爆発曲線に関する数値・数学解析 佐々木多希子 (明大理工)・石渡哲哉 (芝浦工大システム理工) We study a blowup curve for the one dimensional wave equation $\partial_t^2 u \partial_x^2 u = 2^p\partial_t u^p$ with the Dirichlet boundary condition. The purpose of this talk is to show that the blowup curve $T$ satisfies that $T'(x)\rightarrow 1$ as $x\rightarrow 0 + 0\ (1)$ under the suitable initial conditions. To prove the result, we convert the equation into a first order system, and then present some numerical investigations of the blowup curves. From the numerical results, we were able to confirm (1) holds numerically. Moreover, under some assumptions, we were also able to confirm (1) holds mathematically. 

62. 
ある準線形放物型偏微分方程式の後方自己相似解に関する一考察 穴田浩一 (早大高等学院)・石渡哲哉 (芝浦工大システム理工)・牛島健夫 (東京理大理工) In this talk, we consider properties of backward self similar solutions for a quasilinear parabolic equation $v_t=v^{\delta}(v_{xx}+v)$. Their properties are very important to investigate asymptotic behavior of solutions to this parabolic equation, especially, the blowup sets and rates. 

63. 
無限遠ダイナミクスが導く解の爆発レート 松江 要 (九大IMI／九大I2CNER) Geometric treatments of blowup solutions for autonomous ordinary differential equations and their blowup rates are concerned. Our approach focuses on the type of invariant sets at infinity via compactifications of phase spaces, and dynamics on their centerstable manifolds. In particular, we show that dynamics on centerstable manifolds of invariant sets at infinity with appropriate timescale desingularizations as well as blowingup of singularities characterize dynamics of blowup solutions as well as their rigorous blowup rates not only of socalled “typeI” but also other types. 

64. 
速いレートで振る舞う振動爆発解と, 振動発散解 松江 要 (九大IMI／九大I2CNER) Geometric treatments of oscillatory blowup solutions for autonomous ordinary differential equations and their blowup rates are concerned. As in the preceding talk, we apply compactification of phase spaces and timescale desingularization to characterization of blowup solutions. In particular, when divergent solutions are characterized by trajectories on center manifolds of nonhyperbolic periodic orbits on the horizon for desingularized vector fields, they blow up in finite time with infinitely fast oscillation and faster blowup rate than typeI rates in typical cases, while they can also grow up in infinite time in some cases. We see such behavior in a certain system known as Liénard equation. 