±è°ü¼®
|
2019-06-16 16:51:45, Á¶È¸¼ö : 1,118 |
- Download #1 : dff_1_3.jpg (47.8 KB), Download : 1
6. Gauss Curvature II <- Figure I-26 ÂüÁ¶
∘ Definition I-4
𝐤1°ú 𝐤2°¡ °î¸é 𝐌ÀÇ ÇÑ Á¡ 𝐏ÀÇ normal curvature 𝐤n(𝐯)ÀÇ ÃÖ´ë°ª°ú ÃÖ¼Ò°ªÀ̶ó¸é, 𝐤1°ú 𝐤2¸¦ principal curvatures ¶ó
Çϸç ÇØ´ç ¹æÇâÀ» principal directions À̶ó Çϸç, °öÀÎ 𝐊 = 𝐊(𝐏) = 𝐤1𝐤2´Â 𝐏¿¡¼ÀÇ 𝐌ÀÇ Gauss curvature ¶ó ºÎ¸¥´Ù.
∘ Theorem I-5
𝐌ÀÇ ¾î¶² Á¡ 𝐏ÀÇ Gauss curvature : 𝐊(𝐏) = 𝐿/𝑔 ·Î ÁÖ¾îÁø´Ù. <- 𝐿 = det (𝐿𝑖𝑗), 𝑔 = det (𝑔𝑖𝑗)
Proof. ¿ì¸®´Â 𝑔𝑖𝑗v𝑖v𝑗 = 1 ¶ó´Â Á¦ÇÑµÈ ½ÄÀ» extremize(пö·ûù)ÇÔÀ¸·Î½á Áõ¸íÀ» À¯µµÇÏ·Á°í ÇÕ´Ï´Ù.
𝐤 = 𝐤n(𝐯) = 𝐿𝑖𝑗v𝑖v𝑗/𝑔𝑚𝑛v𝑚v𝑛 <- 𝐯: non-zero vectors of 𝐓𝑝𝐌; 𝐯 = v𝑖𝐗𝑖: principal direction vector, ¡Ó𝐤/¡Óv1 = ¡Ó𝐤/¡Óv2 = 0
Let f(x1, x2) = a𝑖𝑗x𝑖x𝑗 <- quadratic form, (a𝑖𝑗): real symmetric n X n matrix: a𝑖𝑗 = a𝑖𝑗, Then ¡Óf/¡Óx𝑟 = 2a𝑟𝑗x𝑗, r = 1,2. ¡Å
[d[f(x)/g(x)]/dx = [f'(x)g(x)- f(x)g'(x)]/g(x)2]¡æ ¡Ó𝐤/¡Óv𝑟 = (2𝑔𝑚𝑛v𝑚v𝑛𝐿𝑟𝑗v𝑗 - 2𝐿𝑖𝑗v𝑖v𝑗𝑔𝑟𝑛v𝑛)/(𝑔𝑚𝑛v𝑚v𝑛)2, r = 1,2
[ºÐÀÚÀÇ 𝐿𝑖𝑗v𝑖v𝑗↦ 𝐤𝑔𝑚𝑛v𝑚v𝑛, ºÐÀÚ ºÐ¸ð¸¦ Á¤¸®]¡æ ¡Ó𝐤/¡Óv𝑟 = [2(𝐿𝑟𝑗 - 𝐤𝑔𝑟𝑗)v𝑗]/𝑔𝑚𝑛v𝑚v𝑛, r = 1,2
At an extreme value we must have (𝐿𝑖𝑗 - 𝐤𝑔𝑖𝑗)v𝑗 = 0, 𝑖 = 1,2. then det(𝐿𝑖𝑗 - 𝐤𝑔𝑖𝑗) = 0. <- v𝑗: non-zero vector; why? [6-24]
𝑔 = det (𝑔𝑖𝑗) = 𝑔11𝑔22 - (𝑔11)2, 𝐿 = 𝐿11𝐿22 - (𝐿11)2; 𝐤2, 𝐤, constant¿¡ ´ëÇØ (𝐤 - 𝐤1)(𝐤 - 𝐤2) = 0 ¹æ½ÄÀ¸·Î Àü°³ÇÕ´Ï´Ù.
𝐤2𝑔 - 𝐤(𝑔11𝐿22 + 𝑔22𝐿11 - 2𝑔12𝐿12) + 𝐿 = 0 or 𝐤2 - 𝐤(𝑔11𝐿22 + 𝑔22𝐿11 - 2𝑔12𝐿12)/𝑔 + 𝐿/𝑔 = 0, ¡Å 𝐊(𝐏) = 𝐤1𝐤2 = 𝐿/𝑔 ▮ *
∘ Gauss curvatureÀÇ µÎ principal directionÀº ¼·Î Á÷±³ÇÑ´Ù.
Proof. Á¡ 𝐏ÀÇ µÎ non-zero principal direction vector¸¦ °¢°¢ 𝐯 = v𝑖𝐗𝑖, 𝐰 = w𝑖𝐗𝑖 ¶ó°í ÇսôÙ.
By Eq. (24), (𝐿𝑖𝑗 - 𝐤1𝑔𝑖𝑗)v𝑗 = 0, (𝐿𝑖𝑗 - 𝐤2𝑔𝑖𝑗)w𝑗 = 0, 𝑖 = 1,2; [𝑖 ⇄ 𝑗, 𝑔𝑖𝑗 = 𝑔𝑗𝑖] ¡æ 𝐿𝑖𝑗v𝑖 = 𝐤1𝑔𝑖𝑗v𝑖, 𝑖 = 1,2 [6-25]
By Eq. (24) (25), (𝐤1 - 𝐤2)𝑔𝑖𝑗v𝑖w 𝑗 = 0. ¡Å𝑔𝑖𝑗v𝑖w 𝑗 = 𝐯 ∙ 𝐰 = 0. ¸¸ÀÏ 𝐤1 ¡Á 𝐤2 À̸é, µÎ principal directionsÀº ¼·Î Á÷±³ÇÑ´Ù.
¸¸ÀÏ 𝐤1 = 𝐤2 À̸é, ¿ì¸®´Â ¼·Î Á÷±³ÇÏ´Â µÎ directionÀ» ¼±ÅÃÇؼ principal directionsÀ̶ó°í ºÎ¸¨´Ï´Ù. ▮
∘ Sphere mapping or Gauss mapping
𝐏 = 𝐗 (u10, u20) ±×¸®°í 𝛺: neighborhood of (u10, u20), 𝐗: one-to-one, continuous inverse 𝐗(𝛺)¡æ 𝛺 ¶ó°í °¡Á¤ÇÑ´Ù¸é,
𝐗(𝛺)ÀÇ °¢ Á¡ 𝐗 (u1, u2)À¸·Î, E3ÀÇ ¿øÁ¡À¸·Î pararell-translated normal vector 𝐔 (u1, u2)¸¦ °ü·Ã½Ãų ¼ö ÀÖ½À´Ï´Ù.
¡«𝐔¡«= 1 À̹ǷÎ, ±× translated vector´Â unit sphere 𝐒2ÀÇ ÇÑÁ¡°ü ¿¬°ü½ÃÄÑ º¼ ¼ö Àִµ¥, ÀÌ·¯ÇÑ 𝐗(𝛺)¡æ 𝐒2 mappingÀ»
Sphere mapping ¶Ç´Â Gauss mapping¶ó ºÎ¸£¸ç. 𝐔(𝛺)´Â 𝐗(𝛺)ÀÇ spherical normal image¶ó ºÎ¸£±âµµ ÇÕ´Ï´Ù.
𝐔´Â sphere¿¡ ¼öÁ÷ÇϹǷÎ, 𝐒2ÀÇ 𝐔(u1, u2)ÀÇ tangent plane 𝐓u𝐒2´Â 𝐗(u1, u2)ÀÇ tangent plane 𝐓x𝐌°ú ÆòÇàÇÕ´Ï´Ù.
¡Å 𝐔 = ∓ (𝐔1 ⨯ 𝐔2) /¡«𝐔1 ⨯ 𝐔2¡« or 𝐔 ∙ 𝐔1 ⨯ 𝐔2 = ∓¡«𝐔1 ⨯ 𝐔2¡«.
∘ Area 𝐔(𝛺) = ∓ ¡ó𝛺¡«𝐔1 ⨯ 𝐔2¡«du1du2 = ¡ó𝛺 𝐔 ∙ 𝐔1 ⨯ 𝐔2 du1du2, 𝐗(𝛺) = ∓ ¡ó𝛺¡«𝐗1 ⨯ 𝐗2¡«du1du2 = ¡ó𝛺 𝐔 ∙ 𝐗1 ⨯ 𝐗2 du1du2
𝐊(𝐏) = lim𝛺 ¡æ (u01, u02) [Area 𝐔(𝛺)/Area 𝐗(𝛺)] [6-26]
Area 𝐔(𝛺) ≈ (𝐔 ∙ 𝐔1 ⨯ 𝐔2)(𝐏) A(𝛺), 𝐗(𝛺) ≈ (𝐔 ∙ 𝐗1 ⨯ 𝐗2)(𝐏) A(𝛺) ¡Å Area 𝐔(𝛺)/𝐗(𝛺) ≈ (𝐔 ∙ 𝐔1 ⨯ 𝐔2 / 𝐔 ∙ 𝐗1 ⨯ 𝐗2) (𝐏)
∘ Lemma (º¸Á¶ Á¤¸®) I-6
𝐔1 ⨯ 𝐔2 = 𝐊 (𝐗1 ⨯ 𝐗2)
Proof. Define the functions 𝐿𝑖𝑗(u1, u2) 𝐿𝑖𝑗 ¡Õ 𝐿𝑗𝑘 𝑔𝑘𝑖, 𝑖,𝑗 =1,2 [6-27]
𝐿𝑖𝑗 𝑔𝑖𝑚 = 𝐿𝑗𝑘 𝑔𝑘𝑖𝑔𝑖𝑚 = 𝐿𝑗𝑘 𝛿𝑘𝑚 = 𝐿𝑗𝑚; 𝐔 ∙ 𝐔 = 1, ¡Å 𝐔 ∙ 𝐔𝑗 = 0; 𝐔𝑗 = a𝑟𝑗 𝐗𝑟, 𝑗 =1,2 <- linear combination of 𝐗1 and 𝐗2
𝐔 ∙ 𝐗𝑘 = 0 ¡æ 𝐔𝑗 ∙ 𝐗𝑘 + 𝐔 ∙ 𝐗𝑗𝑘 = 0 <- differentiated by u𝑗; ¡Å 𝐔𝑗 ∙ 𝐗𝑘 = - 𝐔 ∙ 𝐗𝑗𝑘 = -𝐿𝑗𝑘 = a𝑟𝑗 𝐗𝑟 ∙ 𝐗𝑘 = a𝑟𝑗 𝑔𝑟𝑘, 𝑗,𝑘 = 1,2
<- by eq. 20; [¾çº¯¿¡ 𝑔𝑘𝑖¸¦ °öÇÏ°í sum over k]¡æ - 𝑔𝑘𝑖𝐿𝑗𝑘 = a𝑟𝑗 𝑔𝑟𝑘𝑔𝑘𝑖 = a𝑟𝑗 𝛿𝑖𝑟 = a𝑖𝑗, ¡Å a𝑖𝑗 = - 𝐿𝑖𝑗 <- by eq. 27;
¡Å 𝐔𝑗 = - 𝐿𝑖𝑗 𝐗𝑖, 𝑗 =1,2 <- the equation of Weingarten [6-28]
𝐔1 ⨯ 𝐔2 = (- 𝐿𝑖1𝐗𝑖) ⨯ (- 𝐿𝑘2𝐗𝑘) = (𝐿11𝐗1 + 𝐿21𝐗2) ⨯ (𝐿12𝐗1 + 𝐿22𝐗2) = (𝐿11𝐿22 - 𝐿21𝐿12) (𝐗1 ⨯ 𝐗2) = det (𝐿𝑖𝑗) (𝐗1 ⨯ 𝐗2)
<- 𝐗𝑖 ⨯ 𝐗𝑖 = 𝟎, 𝐗𝑖 ⨯ 𝐗𝑗 = - 𝐗𝑗 ⨯ 𝐗𝑖; [det(𝐀𝐁) = det(𝐀)det(𝐁); det(𝐀-1) = 1/det(𝐀): det(𝐀) is non-zero, iff, 𝐀 is invertible.]
det(𝐿𝑖𝑗) = det(𝐿𝑗𝑘𝑔𝑘𝑖) = det(𝐿𝑗𝑘)det(𝑔𝑘𝑖) = det(𝐿𝑗𝑘)/det(𝑔𝑘𝑖) = 𝐿/𝑔 = 𝐊. ▮
7. Geodesics ** <- Figure I-28 ÂüÁ¶
∘ ÀϹÝÀûÀ¸·Î °î¸é 𝐌 À§ °î¼± 𝛂Àº µÎ°¡Áö ÀÌÀ¯·Î curvature¸¦ °®½À´Ï´Ù. ù°´Â °î¸é ÀÚü°¡ 3Â÷¿øÀÇ °ø°£ ¼Ó¿¡¼ ±ÁÀº °ÍÀ¸·Î
º»ÁúÀûÀ¸·Î´Â ¾Õ¿¡¼ ³íÀÇµÈ normal curvature ÀÔ´Ï´Ù. µÑ°´Â 𝐌ÀÌ °î¸éÀÌµç ¾Æ´Ïµç 𝛂°¡ 𝐌¿¡ ´ëÇØ »ó´ëÀûÀ¸·Î ±ÁÀº Á¤µµ·Î
À̸¦ geodesic curvature ¶ó ºÎ¸£¸ç, ¾Æ·¡¿Í °°ÀÌ Á¤ÀÇÇÕ´Ï´Ù.
∘ 𝛂(s) <- s: arc length; 𝛂' ∙ 𝛂' = 1, ¡Å 𝛂" ∙ 𝛂' = 0; 𝛂" = 𝛂"tan + 𝛂"nor = (u𝑟" + 𝛤𝑟𝑖𝑗 u𝑖'u𝑗')𝐗𝑟 + (𝐿𝑖𝑗 u𝑖'u𝑗')𝐔 <- by eq. (18)
𝛂"tan ∙ 𝐔 = 0 and 𝛂"tan ∙ 𝛂' = (𝛂"tan + 𝛂"nor) ∙ 𝛂' = 𝛂" ∙ 𝛂' = 0 À̹ǷÎ, 𝛂"tan´Â ´ÙÀ½ ½ÄÀÇ the geodesic normal vector ÀÎ,
𝐰 = 𝐔 ⨯ 𝛂' (a vector tangent to 𝐌)¿Í ºñ·ÊÇϴµ¥. ±× propotionality factor¸¦ geodesic curvature ¶ó°í ºÎ¸¨´Ï´Ù.
∘ Definition I-7
𝛂 = 𝛂(s)°¡ s°¡ È£ÀÇ ±æÀÌÀÎ 𝐌ÀÇ °î¼±ÀÏ ¶§, 𝛂(s)¿¡¼ÀÇ geodesic curvature ´Â ´ÙÀ½ ¹æÁ¤½ÄÀÇ ÇÔ¼ö 𝐤g = 𝐤g(s)ÀÌ´Ù.
𝛂"tan = 𝐤g 𝐰 = 𝐤g 𝐔 ⨯ 𝛂' [7-30]
¡Å 𝐤g = 𝐔 ∙ 𝛂' ⨯ 𝛂" <- geodesic curvature 𝐤g = 𝛂"tan∙ 𝐰 = 𝛂" ∙ 𝐰 = 𝛂" ∙ 𝐔 ⨯ 𝛂' <- »ïÁß ³»ÀûÀÇ ¼øȯ [7-31]
∘ Definition I-8
𝛂 = 𝛂(s)°¡ s°¡ È£ÀÇ ±æÀÌÀÎ 𝐌ÀÇ °î¼±ÀÏ ¶§, ¸¸ÀÏ 𝛂°¡ ¸ðµç(every) Á¡¿¡¼ 𝛂"tan = 𝟬 (µ¿µîÇÏ°Ô 𝛂" = 𝛂"nor) À̶ó¸é
𝛂¸¦ ÇϳªÀÇ geodesic À̶ó°í ºÎ¸¥´Ù.
¡Å 𝛂"tan = u𝑟" + 𝛤𝑟𝑖𝑗 u𝑖'u𝑗' = 0, 𝑟 = 1,2, 𝐤g = 𝐔 ∙ 𝛂'⨯ 𝛂" = 0 <- a geodesic, if and onky if, [7-32ab]
∘ Christoffel symbols of the second kind and the first kind
Eq. (17) [𝐗𝑖𝑗 = 𝛤𝑟𝑖𝑗𝐗𝑟 + 𝐿𝑖𝑗𝐔, 𝑖,𝑗 = 1, 2]¿¡¼ Á¤ÀÇµÈ 𝛤𝑟𝑖𝑗¸¦ Christoffel symbols of the second kind ¶ó°í ºÎ¸¨´Ï´Ù.
¶ÇÇÑ 𝛤𝑖𝑗𝑘 = 𝛤𝑟𝑖𝑗𝑔𝑟𝑘 𝑖,𝑗,𝑘 = 1,2 ·Î Á¤ÀÇÇÕ´Ï´Ù. <- 𝛤𝑖𝑗𝑘: Christoffel symbols of the first kind [7-33]
𝐗𝑖𝑗 ∙ 𝐗𝑘 = 𝛤𝑟𝑖𝑗 𝐗𝑟 ∙ 𝐗𝑘 = 𝛤𝑟𝑖𝑗 𝑔𝑟𝑘 = 𝛤𝑖𝑗𝑘 -> ¡Ø 𝛤𝑟𝑖𝑗= 𝛤𝑟𝑗𝑖, 𝛤𝑖𝑗𝑘= 𝛤𝑗𝑖𝑘, 𝛤𝑚𝑖𝑗= 𝛤𝑖𝑗𝑘 𝑔𝑘𝑚 by eq. (17,12) [7-34]
¡Ó𝑔𝑖𝑘/¡Óu𝑗 = (¡Ó /¡Óu𝑗)(𝐗𝑖 ∙ 𝐗𝑘) = 𝐗𝑖𝑗 ∙ 𝐗𝑘 + 𝐗𝑘𝑗 ∙ 𝐗𝑖 = 𝛤𝑖𝑗𝑘 + 𝛤𝑘𝑗𝑖
¡Ó𝑔𝑖𝑘/¡Óu𝑗 = 𝛤𝑖𝑗𝑘 + 𝛤𝑘𝑗𝑖, ¡Ó𝑔𝑗𝑖/¡Óu𝑘 = 𝛤𝑗𝑘𝑖 + 𝛤𝑖𝑘𝑗, ¡Ó𝑔𝑘𝑗/¡Óu𝑖 = 𝛤𝑘𝑖𝑗 + 𝛤𝑗𝑖𝑘 [7-35]
¡Å 𝛤𝑖𝑗𝑘 = 1/2 (¡Ó𝑔𝑖𝑘/¡Óu𝑗 + ¡Ó𝑔𝑘𝑗/¡Óu𝑖 - ¡Ó𝑔𝑗𝑖/¡Óu𝑘) <- Christoffel symbols of the first kind [7-36]
¡Å 𝛤𝑟𝑖𝑗 = (1/2) 𝑔𝑘𝑟 (¡Ó𝑔𝑖𝑘/¡Óu𝑗 + ¡Ó𝑔𝑘𝑗/¡Óu𝑖 - ¡Ó𝑔𝑗𝑖/¡Óu𝑘) <- Christoffel symbols of the second kind [7-37]
Eq.(32a)(37)·Î extrinsicÇÑ Á¤ÀÇÀÇ Christoffel symbol°ú geodesicÀÇ intrinsic ÇعýÀÌ °¡´ÉÇÑ °ÍÀ» ¾Ë ¼ö ÀÖ½À´Ï´Ù!
∘ Christoffel symbols °è»ê <- 𝑔𝑖𝑗 = 𝐗𝑖 ∙ 𝐗𝑗, 𝑔 = det (𝑔𝑖𝑗) = 𝐸𝐺 - 𝐹2, 𝑔𝑖𝑗𝑔𝑗𝑘 = 𝛿𝑘𝑖, 𝑔11= 𝑔22/𝑔, 𝑔12= 𝑔21= -𝑔12/𝑔, 𝑔22= 𝑔11/𝑔
Orthogonal coordinates òÁÎß ñ¨øöͧ: 𝐹 = 𝑔12= 𝑔21= 0, 𝑔12= 𝑔21= 0, 𝑔11= 𝐺/𝐸𝐺 = 1/𝐸 = 1/𝑔11, 𝑔22= 𝐸/𝐸𝐺 = 1/𝐺 = 1/𝑔22
𝛤𝑟𝑖𝑗 = (1/2𝑔𝑟𝑟)(¡Ó𝑔𝑖𝑟/¡Óu𝑗 + ¡Ó𝑔𝑟𝑗/¡Óu𝑖 - ¡Ó𝑔𝑗𝑖/¡Óu𝑟) (no sum) <- replacing 𝑘 by 𝑟 [7-38]
(Case 1) For 𝑗 = 𝑟: 𝛤𝑟𝑖𝑟 = (1/2𝑔𝑟𝑟) ¡Ó𝑔𝑟𝑟/¡Óu𝑖 (no sum) [7-39a]
(Case 2) For 𝑖 = 𝑗 ¡Á 𝑟: 𝛤𝑟𝑖𝑖 = (1/2𝑔𝑟𝑟)(- ¡Ó𝑔𝑖𝑖/¡Óu𝑟) (no sum) <- 𝑔𝑖𝑟= 𝑔𝑗𝑟= 0 [7-39b]
(Case 3) (in dimension > 2) For 𝑖, 𝑗, 𝑟 all distinct: 𝛤𝑟𝑖𝑗 = 0 <- 𝑔𝑖𝑗= 𝑔𝑖𝑟= 𝑔𝑗𝑟= 0 [7-39c]
(in dimension 2) 𝛤111= 𝐸u/2𝐸, 𝛤222= 𝐺v/2𝐺, 𝛤112= 𝛤121= 𝐸v/2𝐸, 𝛤212= 𝛤221= 𝐺u/2𝐺, 𝛤122= -𝐺u/2𝐸<, 𝛤211= -𝐸v/2𝐺 [7-40]
∘ Example 17. Cartesian ÁÂÇ¥°èÀÇ Æò¸é
ds2 = du2 + dv2, 𝐸 = 𝐺 = 1, u"= v"= 0; u = as + b, v = cs + d, ¡Å geodesics: straight lines(òÁàÊ) ▮
∘ Example 18. Geographic(ò¢×â) ÁÂÇ¥°èÀÇ ±¸
𝐗(u, v) = (R cos u cos v, R cos u sin v, R sin v), 𝐸 = R2cos v2, 𝐹 = 0 , 𝐺 = R2,
𝐗1 = (-R sin u cos v, R sin u sin v, 0), 𝐗2 = (-R cos u sin v, R cos u cos v, R cos v),
ds2 = R2cos2v du2 + R2dv2, 𝛤112= 𝛤121= 𝐸v/2𝐸 = -tan v, 𝛤211= -𝐸v/2𝐺 = sin v cos v,
(Geographic ÁÂÇ¥°è ±¸ÀÇ geodesicsÀÌ great circles(ÓÞê)À̶ó´Â Áõ¸íÀº ³ªÁß¿¡ ³ª¿É´Ï´Ù. <- Ex. 12, 14 ÂüÁ¶)
∘ Example 19. Polar(п) ÁÂÇ¥°èÀÇ Æò¸é
𝐗(r, 𝜃) = (rcos𝜃, rsin𝜃, 0), 𝐗1 = (cos𝜃, sin𝜃, 0), 𝐗2 = (-rsin𝜃, rcos𝜃, 0), 𝐸 = 1, 𝐹 = 0, 𝐺 = r2; (u, v)¡æ (r, 𝜃) in Eq. (40)
ds2 = dr2 + r2d𝜃2, 𝛤212= 𝛤221= 1/r, 𝛤122= -r; applying Eq. (32a) [ur" + 𝛤𝑟𝑖𝑗u𝑖'u𝑗' = 0, r=1,2 u1= r, u2= 𝜃]¡æ
d2r/ds2 - r (d𝜃/ds)2 = 0, d2𝜃/ds2 + 2/r (dr/ds)(d𝜃/ds) = 0 [7-41ab]
If we divide Eq. (41b) by 𝜃'= d𝜃/ds, then 1/𝜃'(d𝜃'/ds) + 2/r(dr/ds) = 0, [integrated to s]¡æ ln ∣𝜃'∣ + ln r2 = c.
[expoentiationg]¡æ ∣r2𝜃'∣ = ec or r2 d𝜃/ds = h h: non-zero constant [7-42]
[deviding Eq.(6) by ds2] ¡æ 1 = (dr/ds)2 + r2(d𝜃/ds)2 = (dr/ds)2 + h2/r2 [7-43]
dr/ds = ∓ 1/r(r2 - h2)1/2, [by eq. (40)]¡æ d𝜃/dr = ∓ h/[r(r2-h2)1/2] = ∓ d(cos-1 h/r)/dr
h/r = cos(𝜃 - 𝜃0) <- h: ¿øÁ¡¿¡¼ Á÷¼±±îÁö ¼öÁ÷°Å¸®, 𝜃0: 𝜃 = 0 ¿¡¼ ¿øÁ¡¿¡¼ Á÷¼±±îÁöÀÇ ¼öÁ÷ ±³Â÷Á¡±îÁöÀÇ °¢ [7-44]
±×·¯¹Ç·Î ÀÌ geodesicÀº Á¡ 𝐏(r, 𝜃)¿Í 𝐀(h, 𝜃0)¸¦ ¿¬°áÇÏ´Â Á÷¼± 𝐏𝐀¡êÀÇ ¹æÁ¤½ÄÀÔ´Ï´Ù. <- ¡Ð𝐏𝐎𝐀 = 𝜃 - 𝜃0 ▮
∘ Theorem I-9
𝛂 = 𝛂(s)°¡, a¡Âs¡Âb, 𝐌 À§ÀÇ s°¡ È£ÀÇ ±æÀÌÀÎ °î¼±ÀÏ ¶§, ¸¸ÀÏ 𝛂°¡ µÎ ³¡Á¡À» ¿¬°áÇÏ´Â °î¸é 𝐌 À§ÀÇ °¡Àå ªÀº °î¼±À̶ó¸é,
𝛂´Â ÇÑ geodesicÀÌ´Ù.
Proof. 𝛂(s) = 𝐗(u1(s), u2(s)), length of 𝛂𝜖(s) ≧ length of 𝛂(s),
U𝑖(s, 𝜖) = u𝑖(s) + 𝜖v𝑖(s) for 𝑖=1,2, a¡Âs¡Âb, v𝑖: smooth function v𝑖(a)= v𝑖(b)= 0, 𝑖=1,2, ÀÓÀÇÀÇ (U1, U2) in the domain of 𝐗
For each 𝜖, 𝛂𝜖(s) = 𝐗(U1(s, 𝜖), U1(s, 𝜖)) is a slight variation. [Figure I-28]; ÀÌÁ¦ L(𝜖)¸¦ 𝛂𝜖ÀÇ ±æÀ̶ó°í ÇսôÙ.
L(𝜖) = ¡òab𝜆(s, 𝜖)ds, <- 𝜆(s, 𝜖) = [𝑔𝑖𝑗(U1,U2) ¡ÓU𝑖/¡Ós, ¡ÓU𝑗/¡Ós]1/2; the minimality of L(0) ¡æ L'(0) = ¡òab ¡Ó𝜆(s,0)/¡Ó𝜖 ds = 0
¡Ó𝜆(s,𝜖)/¡Ó𝜖 = 1/2𝜆[(¡Ó𝑔𝑖𝑗)/¡Ó𝜖)(¡ÓU𝑖/¡Ós)(¡ÓU𝑗/¡Ós) + 2𝑔𝑖𝑗(¡ÓU𝑖/¡Ós)(¡Ó2U𝑗/¡Ós¡Ó𝜖)] <- 𝜆=¡î F, 𝜆'=(1/2𝜆)F'; 𝑔𝑖𝑗= 𝑔𝑗𝑖
L'(0) = (1/2)¡òab (¡Ó𝑔𝑖𝑗/¡ÓU𝑘)v𝑘U𝑖'U𝑗' + 2𝑔𝑖𝑘U𝑖'v𝑘') ds = 0 <- 𝜖 = 0, 𝛂0(s) = 𝛂(s), 𝜆(s, 0) = 1; ¡ÓU𝑗/¡Ó𝜖 = v𝑗
°ýÈ£¾È µÑ°Ç×ÀÇ ÀûºÐ: [¡ò udv = uv- ¡ò vdu], u: 2𝑔𝑖𝑘U𝑖', dv: v𝑘'ds, du: ¡Ó/¡Ós[2𝑔𝑖𝑘U𝑖']ds, v = ¡ò dv = ¡ò v𝑘'ds = v𝑘,
uv = 2𝑔𝑖𝑘U𝑖'v𝑘 ∣ba = 0 (¡ñ v𝑘(a) = v𝑘(b) = 0), ¡Å [¡ò udv = - ¡ò vdu = - 2¡Ó(𝑔𝑖𝑘U𝑖'v𝑘)/¡Ós ds
L'(0) = (1/2)¡òab[(¡Ó𝑔𝑖𝑗/¡ÓU𝑘)U𝑖'U𝑗' - 2¡Ó(𝑔𝑖𝑘U𝑖')/¡Ós] v𝑘 ds = 0
¡Å 1/2(¡Ó𝑔𝑖𝑗/¡ÓU𝑘)U𝑖'U𝑗' - ¡Ó(𝑔𝑖𝑘U𝑖')/¡Ós = 0, 𝑘 = 1,2 ¡æ ÀÌÇÏ´Â À̷κÎÅÍ Eq. (32a)¸¦ Ãß·ÐÇÕ´Ï´Ù.
0 = 1/2(¡Ó𝑔𝑖𝑗/¡ÓU𝑘)U𝑖'U𝑗' - (¡Ó𝑔𝑖𝑘/¡ÓU𝑗)U𝑖''U𝑗'- 𝑔𝑚𝑘u𝑚" = [1/2(𝛤𝑖𝑘𝑗 + 𝛤𝑗𝑘𝑖) - (𝛤𝑘𝑗𝑖 + 𝛤𝑖𝑗𝑘)]U𝑖'U𝑗' - 𝑔𝑚𝑘u𝑚"
[𝛤𝑖𝑘𝑗U𝑖'U𝑗' = 𝛤𝑗𝑘𝑖U𝑖'U𝑗'] (dummi index ±³È¯); 𝛤𝑗𝑘𝑖 = 𝛤𝑘𝑗𝑖; ¡Å 𝑔𝑚𝑘u𝑚" + 𝛤𝑖𝑗𝑘U𝑖'U𝑗' = 0; ¾çº¯¿¡ 𝑔𝑘r °öÇÏ°í k¿¡ ´ëÇØ ÇÕ»êÀ» Çϸé,
𝑔𝑚𝑘𝑔𝑘ru𝑚" + 𝛤𝑖𝑗𝑘𝑔𝑘rU𝑖'U𝑗' = 0, 𝜎𝑚ru𝑚" + 𝛤 r 𝑖𝑗U𝑖'U𝑗' = 0, ur" + 𝛤 r 𝑖𝑗U𝑖'U𝑗' = 0, r = 1,2; ¡Å 𝛂´Â ÇϳªÀÇ geodesicÀÔ´Ï´Ù. ▮
∘ Theorem I-10
𝐌ÀÇ Á¡ 𝐏¿Í ±× °÷ÀÇ unit tangent vector 𝐯°¡ ÁÖ¾îÁö¸é, 𝛂(0) = 𝐏 ±×¸®°í 𝛂'(0) = 𝐯ÀÎ À¯ÀÏÇÑ geodesic 𝛂(s)°¡ Á¸ÀçÇÑ´Ù.
Proof. 𝐏 = 𝐗(u01, u01), 𝐯 = v𝑖𝐗𝑖(u01, u01) À̶ó°í Çϸé geodesicÀÇ Á¤ÀÇ Eq. (32a)¿¡ µû¶ó À¯ÀÏÇÑ function ur(s)´Â,
u𝑟" + 𝛤𝑟𝑖𝑗 u𝑖'u𝑗' = 0, ur = ur0, ur'(0) = v' for r = 1,2 <- 𝛂(s) = 𝐗(u1(s), u2(s)), s: arc length
f(s) = ¡«𝛂'(s)¡«2 = 𝑔𝑖𝑗 u𝑖'u𝑗' = constant C <- ds2 by eq. (6), the first fundamental form; f(0): unit length
f'(s) = (¡Ó𝑔𝑖𝑗/¡Óu𝑘)u𝑖'u𝑗'u𝑘' + 𝑔𝑖𝑗 u𝑖"u𝑗' + 𝑔𝑖𝑗 u𝑖'u𝑗" = (𝑔𝑗𝑟𝛤𝑟𝑖𝑘 + 𝑔𝑖𝑟𝛤𝑟𝑗𝑘)u𝑖'u𝑗'u𝑘' + 𝑔𝑖𝑗 u𝑖"u𝑗' + 𝑔𝑖𝑗 u𝑖'u𝑗"
= 𝑔𝑖𝑟u𝑖'(u𝑟" + 𝛤𝑟𝑗𝑘 u𝑗'u𝑘') + 𝑔𝑟𝑗u𝑗'(u𝑟" + 𝛤𝑟𝑖𝑘 u𝑖'u𝑘') = 0, and so f(s) ¡Õ 1. <- unit tangent vector
Áï, Geodesic ÇÊ¿äÃæºÐÁ¶°Ç½Ä Eq. (32a)À» Àû¿ëÇϸé metric formÀÌ ¼º¸³ÇÏ°í, ±× ¿ªµµ ¼º¸³ÇϹǷΠÁõ¸íÀÌ ¿Ï·áµË´Ï´Ù. ▮
∘ Exercise I-7-12. Sppose 𝐌 has metric form ds2 = 𝐸 du2 + 𝐺 dv2 with 𝐸v = 𝐺v = 0.
(a) Verify that the only non-zero Christoffel symbol 𝛤𝑟𝑖𝑗 are 𝛤111 = 𝐸u /2𝐸, 𝛤212 = 𝛤221 = 𝐺u /2𝐺, 𝛤122 = - 𝐺u /2𝐸. Solution. Eq. (40)
(b) Show that a geodesic on 𝐌 satisfies v" + 𝐺u /𝐺 u' v' = 0 and intergrate this equation to obtain 𝐺v' = h (a non-zero constant)
Solution. u𝑟" + 𝛤𝑟𝑖𝑗 u𝑖'u𝑗' = 0, 𝑟 = 1,2, Hence v " + 2𝛤212u'v' = 0. ¡Å v" + 𝐺u /𝐺 u'v' = 0. Again, d(𝐺v')/𝐺ds = v" + (𝐺u /𝐺)u'v' = 0.
Therefore d(𝐺v')/ds = 0, and 𝐺v' = h, h = non-zero constant.
(c) Refering to Example 19, combine 𝐺v' = h with 𝐸 (u')2 + 𝐺 (v')2 = 1 to obtain dv/du = ∓ h¡î 𝐸 /{¡î 𝐺 ¡î (𝐺 - h2)}. [7-45]
Solution. dv/dh = h/𝐺. 1 = 𝐸(du/ds)2 + 𝐺(du/ds)2= 𝐸(du/ds)2 + 𝐺(h/𝐺)2= 𝐸(du/ds)2 + h2/𝐺. Hence du/ds = ∓ ¡î {(1 - h2/𝐺)/𝐸},
dv/du = ∓ h/𝐺 ¡î {E/(1 - h2/𝐺)} = ∓ h¡î 𝐸 /{¡î 𝐺 ¡î (𝐺 - h2)}. ▮
∘ Exercise I-7-14. If 𝐌 has metric ds2 = 𝐸 du2 + 𝐺 dv2 with 𝐸u = 𝐺u = 0. a geodesic on 𝐌 satisfies du/dv = ∓ h¡î 𝐺 /¡î 𝐸 ¡î (𝐸 - h2).
(A proof is immediately obtained if in Exercise 12 we Intechange u with v, 𝐸 with 𝐺, and 1 with 2.)
(a) Show that a geodesic on the sphere satisfies du/dv = h sec2v /¡î (R2 - h2 sec2v) = h sec2v /¡î (R2 - h2 - h2 tan2v). h: const.
Solution. du/dv = ∓ h¡î 𝐺 /{¡î 𝐸 ¡î (𝐸 - h2)} 𝐸u' = h, 𝐸 = R2 cos2 v, 𝐺 = R2,
du/dv = ∓ h¡î 𝐺 /¡î 𝐸 ¡î (𝐸 - h2) = hR / (R cos v ¡î (R2 cos v - h2) = h sec2v /¡î (R2 - h2 sec2v) = h sec2v /¡î (R2 - h2 - h2 tan2v).
(b) By means of the substitution w = h tan v, intergrate the above equation to obtain cos(u - u0) + 𝛾 tan v = 0, u0), 𝛾 = constants. [46]
[Use ¡ò (a2 - x2)-1/2 dx = - cos-1 (x/a) + c.]
Solution. dw = h sec2 v dv, u = ¡ò {h sec2 v / ¡î (r2 - h2 - h2 tan2} dv = - ¡ò {-1/¡î R2 - h2 - w2)} dw = - cos-1(w /¡î {R2 - h2)} + u0.
Therefore cos (u - u0) = w /¡î {R2 - h2) = h tan v /¡î {R2 - h2). Let 𝛾 = -h /¡î {R2 - h2) be. Then we have cos (u - u0) + 𝛾 tan v = 0.
(c) Show that Eq. (46), when reexpressed in terms of Cartesian coordinates, x = R cos u cos v, y = R sin u cos v, z = R sin v
is a linear equation of the form 𝛼x + 𝛽y + 𝛾z = 0, and so represednts a plane pf the sphere through the orgin, i.e., a great circle.
Solution. By multiplying R cos v, R cos v cos (u - u0) + 𝛾 cos v tan v = 0, R cos v (cos u cos u0 + sin u sin u0) + 𝛾 sin v = 0.
Hence if 𝛼 = cos u0, 𝛽 = sin u0, then we have 𝛼x + 𝛽y + 𝛾z = 0. (ÀÌ´Â 𝐧 (𝛼, 𝛽, 𝛾)°ú Á÷±³ÇÏ¸é¼ ¿øÁ¡À» Áö³ª´Â Æò¸é 𝐏 (z,y,z)ÀÔ´Ï´Ù.
Áï, 𝐧 ∙ 𝐏 = 0. µû¶ó¼ Áö¸®Àû(geographical) ÁÂÇ¥°èÀÇ geodesicÀº ´ë¿ø(great circle)ÀÇ ÀϺÎÀÔ´Ï´Ù.) ▮
p.s. determinant´Â ¼±Çüµ¶¸³ ¿©ºÎ, ¿ªÇà·Ä Á¸Àç ¿©ºÎ, »ïÁß³»ÀûÀÇ ºÎÇÇ ¹× Gauss curvature »êÁ¤ÀÇ '°áÁ¤ÀÚ'-google-ÀÓ. ('Çà·Ä½Ä'Àº ºñÃßõ)
* ▮ 'the proof is complete ' Áõ¸íÀÌ ¿Ï·áµÊÀ» ³ªÅ¸³»´Â ±âÈ£
** geodesic: ´ëÇѼöÇÐȸÀÇ ÃøÁö¼±(ö´ò¢àÊ), ¹°¸®ÇÐȸÀÇ Áö¸§±æÀε¥, ¿©±â¼´Â ÃøÁö¼±/¿µ¾î·Î Ç¥±â |
|
|