• 沒有找到結果。

熱帶幾何與差分方程之研究

N/A
N/A
Protected

Academic year: 2021

Share "熱帶幾何與差分方程之研究"

Copied!
16
0
0

加載中.... (立即查看全文)

全文

(1)

行政院國家科學委員會專題研究計畫 成果報告

熱帶幾何與差分方程之研究

研究成果報告(精簡版)

計 畫 類 別 : 個別型

計 畫 編 號 : NSC 99-2115-M-004-003-

執 行 期 間 : 99 年 08 月 01 日至 100 年 09 月 30 日

執 行 單 位 : 國立政治大學應用數學學系

計 畫 主 持 人 : 蔡炎龍

報 告 附 件 : 國外研究心得報告

公 開 資 訊 : 本計畫可公開查詢

中 華 民 國 100 年 12 月 28 日

(2)

中 文 摘 要 : 本計畫研究熱帶幾何於差分方程之應用。差分方程的解自然

是熱帶亞純函數, 因此我們討論了熱帶亞純函數的基本定義

及性質。接著我們研究解熱帶線性方程式的方法。最後我們

應用這些基礎解出一些差分方程式。

中文關鍵詞: 熱帶幾何, 熱帶亞純函數, 差分方程

英 文 摘 要 : The project study tropical geometry objects arising

from difference equations. The solutions to

difference equations are naturally tropical

meromorphic functions. Therefore, we study the

properties of as well as techniques that can apply to

tropical meromorphic functions. As the classical

situation, solving linear systems are essential to

solving difference equations. Hence we find explicit

algorithm to solve tropical linear systems. Finally,

we solve some explicit examples of difference

equations with the technique we provide.

英文關鍵詞: tropical geometry, tropical meromorphic function,

difference equation

(3)

國科會專題研究成果報告:

熱帶幾何與差分方程之研究

計畫編號: NSC 992115M004 003

-計畫

主持人: 蔡炎龍 (政治大學應用數學系)

(4)

Chapter 1

報告

告內

內容

1.1

Introduction

The project study tropical geometry objects arising from difference equations. The solutions to difference equations are naturally tropical meromorphic func-tions. Therefore, we study the properties of as well as techniques that can apply to tropical meromorphic functions. As the classical situation, solving linear systems are essential to solving difference equations. Hence we find ex-plicit algorithm to solve tropical linear systems. Finally, we solve some exex-plicit examples of difference equations with the technique we provide.

While checking if an arbitrary Fano hypersurface K-stable or not, tropical polynomials come out naturally. It is interesting to see if we can apply our techniques to study K-stability problems.

1.2

Tropical Meromorphic Functions

The tropical semi-ring is the set R ∪ {−∞} with addition and multiplication defined as followings.

Definition 1 (tropical semiring). A tropical semiring is (T, ⊕, ), where T = R ∪ {−∞}. For any x, y ∈ T, define

• x ⊕ y := max{x, y} • x y := x + y

One can then of course define tropical polynomials: they are just classical polynomials with tropical operations. More precisely, let x = (x1, x2, . . . , xn) and i = (i1, i2, . . . , in), we denote x i1 1 x

i2

2 · · · x in n by x i. A tropical polynomial is of the form

f (x) = ⊕ X

i∈I

(5)

where I is a finite subset of (N ∪ {0})n.

A tropical hypersurface is of course the zero locus (roots) of a tropical poly-nomial. Notice that a root of a tropical polynomial is actually a point where the function fails to be linear.

Definition 2 (Tropical Hypersurface). Let f (x) be a tropical polynomial. Eval-uate this polynomial is to find the maximum of the linear forms αi+ hx, ii. A point in the tropical hypersurface Hfis the maximum of the linear forms achieve at least twice and it is exactly where the graph fails to be linear.

Motivated by [4, 5], we give the following definitions to tropical meromorphic functions.

Definition 3 (Tropical Meromorphic Function). We say that a function f is a tropical meromorphic function if

(a) f is a continuous, piecewise linear function with integer slopes on R, and (b) there exist x0∈ R such that f0(x0) is constant for all x < x0.

Definition 4 (R-Tropical Meromorphic Function). We say that a function f is a R-tropical meromorphic function if f is a continuous, piecewise linear function with integer slopes on R.

Definition 5 (Extended Tropical Meromorphic Function). We say that a func-tion f is a extended tropical meromorphic funcfunc-tion if

(a) f is a continuous, piecewise linear function on R, and (b) there exist x0∈ R such that f0(x0) is constant for all x < x0.

Definition 6 (Extended R-Tropical Meromorphic Function). We say that a function f is a extended R-tropical meromorphic function if f is a continuous, piecewise linear function on R.

While solving difference equations, one can choose one of the above defini-tions to fit the needs. We refer to our paper [8] for motivadefini-tions for this definition and also some important properties of tropical meromorphic functions.

1.3

Tropical Linear Systems

A tropical linear system is a linear system with tropical operators. Solving tropical linear systems has been discussed in [2, 3, 6]. We survey and extend these results, and be able to solve all tropical linear systems of the following forms.

T1: A x = 0 T2: A x = b

(6)

T3: A x ≤ b T4: A x = B x T5: A x ≤ B x T6: C x = D y

T7: A x ⊕ a = B x ⊕ b

The techniques of solving these linear systems are different, yet some of them are related to each other. For instance, linear systems of type T1 and T2, are basically the same. Type T4, T5, T6, T7 can be regarded as the same type of problems. Finally, type T3 is a general case of type T2 in some sense. We have all details in [9], where one can find all algorithms to solve these linear systems. For illustrating the idea, we give an example here.

Example 1. Consider the tropical linear system A x = b, where

A =   3 7 −1 −∞ 6 7 −∞ −∞ 1 0 1 −∞  , b =   15 18 13  。 Write x =     x1 x2 x3 x4    

. Since the entries of 4th column A(4) of A are all −∞, so x4 can be anything. For instance,

A(1) x =3 7 −1 −∞     x1 x2 x3 x4     = 15 implies that max{3 + x1, 7 + x2, −1 + x3, −∞ + x4} = 15.

Since −∞ + x4 = −∞ can not be the maximum value, so we can ignore this term. Hence the linear system we need to solve is the following:

  3 7 −1 6 7 −∞ 1 0 1     x1 x2 x3  =   15 18 13  . Now, since the first row of A multiply by x is 15, we get

A(1)   x1 x2 x3  = 15, which is max{3 + x1, 7 + x2, −1 + x3} = 15.

(7)

Subtracting all terms by 15, we get an equivalent expression: max{−12 + x1, −8 + x2, −16 + x3} = 0. That is, −12 −8 −16   x1 x2 x3  = 0. Do the same thing to each raw, we obtain

  −12 −8 −16 −12 −11 −∞ −12 −13 −12     x1 x2 x3  =   0 0 0  .

For each column of A, find the maximum entry, and change every other entries to −∞. The solutions to the new linear system will be the same as before. The new linear system becomes:

  −12 −8 −∞ −12 −∞ −∞ −12 −∞ −12     x1 x2 x3  =   0 0 0  . Finally, we can solve the linear system:

x =     12 x2 x3 x4     , where    x2≤ 8 x3≤ 12 x4∈ T .

1.4

Applications to Difference Equations

Consider ultradiscrete equations

y(x + 1) = y(x) c= cy(x) (c ∈ R). (1.1) which were discussed in [4] and [5], and we want to find the solutions y(x) which are tropical meromorphic functions. To solve the equation, we consider certain special tropical meromorphic functions introduced by Laine and Tohge [5]:

eα(x) := α[x](x − [x]) + [x]−1 X j=−∞ αj= α[x](x − [x] + 1 α − 1), where α is a real number with |α| > 1. Moreover, they define eβ(x) := ∞ X j=[x] βj−β[x](x−[x]) = ∞ X j=[x]+1 βj+β[x](1−x+[x]) = β[x]( 1 1 − β−x+[x]), where β is real number with |β| < 1.

In [1], we redefine eα(x), which still has similar properties as the original one, and will be easier to use in certain cases.

(8)

Figure 1.1: e2(x) is an aproximate function of 21+x.

Example 2. Figure 1.1 shows the graph of e2(x). Now, let’s consider

Halburd and Southall [4] show that equation (1.1) admits a nonconstant tropical meromorphic soluction on R if and only if c = ±1. We gave a different proof in [1].

Theorem 1. The equation (1.1) admits a nonconstant tropical meromorphic soluction on R if and only if c = ±1.

If we allow the solutions to be an extend tropical meromorphic function, then we have the following theorem.

Theorem 2.

y(x + 1) = cy(x), c 6= 0, ±1. (1.2) Given an arbitrary extended tropical meromorphic solution f to equation (1.2) with discontinuities of slope at x1, x2, · · · , xk in [0, 1), then f can be represented as a linear combination of finite shifts of the function e2(x), that is,

f (x) = k X i=1 wf(xi)e2(x − xi) = k X i=1 wf(xi)e(x − xi)

Now, let’s consider the equation

y(x + 1) = y(x) + b (b ∈ R) (1.3) It is clear that y(x) = bx is a solution function of equation (1.3), and if h(x) is an extended tropical meromorphic function such that h(x + 1) = h(x), then h(x) + bx is an extended tropical meromorphic solution function of equation

(9)

(1.3). Moreover, all solution functions of equation (1.3) is the form h(x) + bx, where h(x) is a tropical 1-periodic function.

And now the equation

y(x + 1) = ay(x) + b (a, b ∈ R, a 6= 0, ±1) (1.4) can be turned into

y(x + 1) − b

1 − a = a(y(x) − b

1 − a) (1.5) Let z(x) = y(x) − b

1 − a, equation (1.5) turn into

z(x + 1) = az(x) (1.6) By Theorem 2, the solution function of equation (1.6) can be represented as z(x) =Pk

i=1aie(x − xi), where 0 ≤ xi < 1, ai ∈ R \ {0}. Hence, the solution function of equation (1.4) can be represented as

y(x) = ( k X i=1 aie(x − xi)) + b 1 − a. Last case, consider the equation

y(x + 1) = −y(x) + b (b ∈ R). (1.7) Subtract both sides by −b

2, it turns into y(x + 1) −b

2 = −(y(x) − b 2)

Hence, by Theorem 1, the solution function of equation (1.7) can be represented as

y(x) = u(x + 1) − u(x) + b 2

where u(x) is any 2-periodic tropical meromorphic function. So far we have discussed all the circumstances of equation y(x + 1) = ay(x) + b.

For the example of second order difference equations, one can see [5], Laine and Tohge consider equation

y(x + 1) + y(x − 1) = cy(x) for c ∈ R.

(10)

Chapter 2

成果

果自

自評

In this project, we study the properties of tropical meromorphic functions and published a journal paper [8]. Moreover, two graduate students wrote the meth-ods of solving linear systems and solving certain difference equations as their mater’s thesis [1, 9].

While visiting the University of California, Irvine, we found that the tech-niques we had might help to deal with K-stablility problem for Fano hypersur-faces, since while checking K-stability of Fano hypersurhypersur-faces, there naturally comes out some tropical polynomials [7]. We will continue to study the relation between K-stability and tropical geometry in the following years.

(11)

Bibliography

[1] Liang Chen. Tropical meromorphic functions and their application on dif-ference equations. Master’s thesis, National Chengchi University, 2011. [2] Guy Cohen, Jean P. Quadrat, Geert J. Olsder, and Fran¸cois Baccelli.

Syn-chronization and linearity, an algebra for discrete event systems, 1992. [3] R. A. Cuninghame-Green and P. Butkovic. The equation a ⊗ x = b ⊗ y

over (max, +). Theor. Comput. Sci., 293:3–12, February 2003.

[4] Rodney G. Halburd and Neil J. Southall. Tropical Nevanlinna theory and ultradiscrete equations. Int. Math. Res. Not. IMRN, (5):887–911, 2009. [5] Ilpo Laine and Kazuya Tohge. Tropical Nevanlinna theory and second main

theorem. Proc. Lond. Math. Soc. (3), 102(5):883–922, 2011.

[6] E. Lorenzo and M. J. de la Puente. An algorithm to describe the solution set of any tropical linear system A⊙x=B⊙xA⊙x=B⊙x. Linear Algebra and its Applications, March 2011.

[7] Zhiqin Lu. K energy and k stability on hypersurfaces. COMMUNICATIONS IN ANALYSIS AND GEOMETRY, 12:599, 2004.

[8] Yen lung Tsai. Working with one-variable tropical meromorphic functions. to appear in Taiwanese Journal of Mathematics.

[9] Jun-Bo You. On tropical linear systems. Master’s thesis, National Chengchi University, 2012.

(12)

國科會專題研究出國報告:

熱帶幾何與差分方程之研究

計畫編號: NSC 992115M004 003

-蔡

炎龍

問地點: 美國加州大學爾灣校區

問時間: 2011 年 8 月 30 日至 2011 年 9 月 8 日

非常感謝國科會的支援, 本人於 2011 年暑假得已到美國加州大學爾灣校區 (University of

California, Irvine) 數

學系訪問 Lu, Zhiqin 教授。這次訪問的目的主要是和 Lu 教授討論

K-Stability 理論

熱帶化的可能性。

複幾何關注 K¨ahler-Einstein metric 的存在性, 如果 first Chern class 為 0 時, Shing-Tung

Yau 證明必有 K¨

ahler-Einstein metric, 而 Thierry Aubin

和 Shing-Tung Yau 又證明了 first

Chern class 小於 0 時也有。因此問題就在 first Chern class 大於 0 的情況。

在 Lu 教授的 “E Energy and K Stability on Hypersurface,” 考慮在 CP

n

中的 Fano

hyper-surface M ,

也就是 first Chern class 大於 0 的 hypersurface。

我們考慮的 Fano hypersurface M, 必需是 K-stable, 才有可能有 K¨ahler-Einstein metric。而

在計算的過程中, 有一個計算事實上就是定義 M 的多項式熱帶化的解果。這是一個意外還是理所

當然的, 是一個有意思的問題。更重要的是, 如果將來確認了這個過程, 是不是能推廣到 complete

intersection 的

情況? 這是我們希望再繼續深入的課題。

(13)

國科會補助計畫衍生研發成果推廣資料表

日期:2011/12/28

國科會補助計畫

計畫名稱: 熱帶幾何與差分方程之研究 計畫主持人: 蔡炎龍 計畫編號: 99-2115-M-004-003- 學門領域: 幾何與拓樸學

無研發成果推廣資料

(14)

99 年度專題研究計畫研究成果彙整表

計畫主持人:

蔡炎龍

計畫編號:

99-2115-M-004-003-計畫名稱:

熱帶幾何與差分方程之研究

量化

成果項目

實際已達成

數(被接受

或已發表)

預期總達成

數(含實際已

達成數)

本計畫實

際貢獻百

分比

單位

備 註

質 化 說

明:如 數 個 計 畫

共 同 成 果、成 果

列 為 該 期 刊 之

封 面 故 事 ...

期刊論文

1

1

100%

SCI 期刊乙篇

研究報告/技術報告

3

3

100%

國 科 會 研 究 報 告

乙篇、碩士論文兩

研討會論文

1

1

100%

中正大學代數/幾

何研討會

論文著作

專書

0

0

100%

申請中件數

0

0

100%

專利

已獲得件數

0

0

100%

件數

0

0

100%

件 無

技術移轉

權利金

0

0

100%

千元 無

碩士生

3

3

100%

碩 士 班 參 與 學 生

博士生

0

0

100%

博士後研究員

0

0

100%

國內

參與計畫人力

(本國籍)

專任助理

0

0

100%

人次

期刊論文

0

0

100%

研究報告/技術報告

0

0

100%

研討會論文

0

0

100%

論文著作

專書

0

0

100%

章/本 無

申請中件數

0

0

100%

專利

已獲得件數

0

0

100%

件數

0

0

100%

件 無

技術移轉

權利金

0

0

100%

千元 無

碩士生

0

0

100%

博士生

0

0

100%

博士後研究員

0

0

100%

國外

參與計畫人力

(外國籍)

專任助理

0

0

100%

人次

(15)

其他成果

(

無法以量化表達之成

果如辦理學術活動、獲

得獎項、重要國際合

作、研究成果國際影響

力及其他協助產業技

術發展之具體效益事

項等,請以文字敘述填

列。)

固訂每週舉行相關研討。

成果項目

量化

名稱或內容性質簡述

測驗工具(含質性與量性)

0

課程/模組

0

電腦及網路系統或工具

0

教材

0

舉辦之活動/競賽

0

研討會/工作坊

0

電子報、網站

0

目 計畫成果推廣之參與(閱聽)人數

0

(16)

國科會補助專題研究計畫成果報告自評表

請就研究內容與原計畫相符程度、達成預期目標情況、研究成果之學術或應用價

值(簡要敘述成果所代表之意義、價值、影響或進一步發展之可能性)

、是否適

合在學術期刊發表或申請專利、主要發現或其他有關價值等,作一綜合評估。

1. 請就研究內容與原計畫相符程度、達成預期目標情況作一綜合評估

■達成目標

□未達成目標(請說明,以 100 字為限)

□實驗失敗

□因故實驗中斷

□其他原因

說明:

2. 研究成果在學術期刊發表或申請專利等情形:

論文:■已發表 □未發表之文稿 □撰寫中 □無

專利:□已獲得 □申請中 ■無

技轉:□已技轉 □洽談中 ■無

其他:(以 100 字為限)

完成期刊論文乙篇, 碩士論文兩篇。

3. 請依學術成就、技術創新、社會影響等方面,評估研究成果之學術或應用價

值(簡要敘述成果所代表之意義、價值、影響或進一步發展之可能性)(以

500 字為限)

已完成基礎研究工作, 未來可望應用於更深入, 如 Fano Varieties 之 K-Stability 等

問題。

數據

Figure 1.1: e 2 (x) is an aproximate function of 2 1+x .

參考文獻

相關文件

In this Learning Unit, students are required to solve compound linear inequalities in one unknown involving logical connectives “and” or “or”, quadratic inequalities in one

化成 reduced echelon form 後由於每一個 row 除了該 row 的 pivot 外, 只剩 free variables (其他的 pivot variable 所在的 entry 皆為 0), 所以可以很快地看出解的形式..

Wang, A recurrent neural network for solving nonlinear convex programs subject to linear constraints, IEEE Transactions on Neural Networks, vol..

Abstract Based on a class of smoothing approximations to projection function onto second-order cone, an approximate lower order penalty approach for solving second-order cone

Based on a class of smoothing approximations to projection function onto second-order cone, an approximate lower order penalty approach for solving second-order cone

where L is lower triangular and U is upper triangular, then the operation counts can be reduced to O(2n 2 )!.. The results are shown in the following table... 113) in

This kind of algorithm has also been a powerful tool for solving many other optimization problems, including symmetric cone complementarity problems [15, 16, 20–22], symmetric

Since the subsequent steps of Gaussian elimination mimic the first, except for being applied to submatrices of smaller size, it suffices to conclude that Gaussian elimination