TOPICS
Search

Geodesic


A geodesic is a locally length-minimizing curve. Equivalently, it is a path that a particle which is not accelerating would follow. In the plane, the geodesics are straight lines. On the sphere, the geodesics are great circles (like the equator). The geodesics in a space depend on the Riemannian metric, which affects the notions of distance and acceleration.

Geodesics preserve a direction on a surface (Tietze 1965, pp. 26-27) and have many other interesting properties. The normal vector to any point of a geodesic arc lies along the normal to a surface at that point (Weinstock 1974, p. 65).

Furthermore, no matter how badly a sphere is distorted, there exist an infinite number of closed geodesics on it. This general result, demonstrated in the early 1990s, extended earlier work by Birkhoff, who proved in 1917 that there exists at least one closed geodesic on a distorted sphere, and Lyusternik and Schnirelmann, who proved in 1923 that there exist at least three closed geodesics on such a sphere (Cipra 1993, p. 28).

For a surface given parametrically by x=x(u,v), y=y(u,v), and z=z(u,v), the geodesic can be found by minimizing the arc length

 I=intds=intsqrt(dx^2+dy^2+dz^2).
(1)

But

dx=(partialx)/(partialu)du+(partialx)/(partialv)dv
(2)
dx^2=((partialx)/(partialu))^2du^2+2(partialx)/(partialu)(partialx)/(partialv)dudv+((partialx)/(partialv))^2dv^2,
(3)

and similarly for dy^2 and dz^2. Plugging in,

 I=int{[((partialx)/(partialu))^2+((partialy)/(partialu))^2+((partialz)/(partialu))^2]du^2+2[(partialx)/(partialu)(partialx)/(partialv)+(partialy)/(partialu)(partialy)/(partialv)+(partialz)/(partialu)(partialz)/(partialv)]dudv+[((partialx)/(partialv))^2+((partialy)/(partialv))^2+((partialz)/(partialv))^2]dv^2}^(1/2).
(4)

This can be rewritten as

I=intsqrt(P+2Qv^'+Rv^('2))du
(5)
=intsqrt(Pu^('2)+2Qu^'+R)dv,
(6)

where

v^'=(dv)/(du)
(7)
u^'=(du)/(dv)
(8)

and

P=((partialx)/(partialu))^2+((partialy)/(partialu))^2+((partialz)/(partialu))^2
(9)
Q=(partialx)/(partialu)(partialx)/(partialv)+(partialy)/(partialu)(partialy)/(partialv)+(partialz)/(partialu)(partialz)/(partialv)
(10)
R=((partialx)/(partialv))^2+((partialy)/(partialv))^2+((partialz)/(partialv))^2.
(11)

Starting with equation (◇)

I=intsqrt(P+2Qv^'+Rv^('2))du
(12)
=intLdu,
(13)

and taking derivatives,

(partialL)/(partialv)=1/2(P+2Qv^'+Rv^('2))^(-1/2)((partialP)/(partialv)+2(partialQ)/(partialv)v^'+(partialR)/(partialv)v^('2))
(14)
(partialL)/(partialv^')=1/2(P+2Qv^'+Rv^('2))^(-1/2)(2Q+2Rv^'),
(15)

so the Euler-Lagrange differential equation then gives

 ((partialP)/(partialv)+2v^'(partialQ)/(partialv)+v^('2)(partialR)/(partialv))/(2sqrt(P+2Qv^'+Rv^('2)))-d/(du)((Q+Rv^')/(sqrt(P+2Qv^'+Rv^('2))))=0.
(16)

In the special case when P, Q, and R are explicit functions of u only,

 (Q+Rv^')/(sqrt(P+2Qv^'+Rv^('2)))=c_1
(17)
 (Q^2+2QRv^'+R^2v^('2))/(P+2Qv^'+Rv^('2))=c_1^2
(18)
 v^('2)R(R-c_1^2)+2v^'Q(R-c_1^2)+(Q^2-Pc_1^2)=0
(19)
 v^'=1/(2R(R-c_1^2))[2Q(c_1^2-R)+/-sqrt(4Q^2(R-c_1^2)^2-4R(R-c_1^2)(Q^2-Pc_1^2))].
(20)

Now, if P and R are explicit functions of u only and Q=0,

 v^'=(sqrt(4R(R-c_1^2)Pc_1^2))/(2R(R-c_1^2))=c_1sqrt(P/(R(R-c_1^2))),
(21)

so

 v=c_1intsqrt(P/(R(R-c_1^2)))du.
(22)

In the case Q=0 where P and R are explicit functions of v only, then

 ((partialP)/(partialv)+v^('2)(partialR)/(partialv))/(2sqrt(P+Rv^('2)))-d/(du)((Rv^')/(sqrt(P+Rv^('2))))=0,
(23)

so

 (partialP)/(partialv)+v^('2)(partialR)/(partialv)-2sqrt(P+Rv^('2))R[(v^(''))/(sqrt(P+Rv^('2)))+(-1/2)(v^'(2Rv^'v^('')))/((P+Rv^('2))^(3/2))]=0
(24)
 (partialP)/(partialv)+v^('2)(partialR)/(partialv)-2Rv^('')+(2R^2v^('2)v^(''))/(P+Rv^('2))=0
(25)
 (Rv^('2))/(sqrt(P+Rv^('2)))-sqrt(P+Rv^('2))=c_1
(26)
 Rv^('2)-(P+Rv^('2))=c_1sqrt(P+Rv^('2))
(27)
 (-P/(c_1))^2=P+Rv^('2)
(28)
 (P^2-c_1^2P)/(Rc_1^2)=v^('2),
(29)

and

 u=c_1intsqrt(R/(P^2-c_1^2P))dv.
(30)

For a surface of revolution in which y=g(x) is rotated about the x-axis so that the equation of the surface is

 y^2+z^2=g^2(x),
(31)

the surface can be parameterized by

x=u
(32)
y=g(u)cosv
(33)
z=g(u)sinv.
(34)

The equation of the geodesics is then

 v=c_1int(sqrt(1+[g^'(u)]^2)du)/(g(u)sqrt([g(u)]^2-c_1^2)).
(35)

See also

Blaschke Conjecture, Ellipsoid Geodesic, Geodesic Curvature, Geodesic Dome, Geodesic Equation, Geodesic Mapping, Geodesic Triangle, Graph Geodesic, Great Circle, Harmonic Map, Oblate Spheroid Geodesic, Paraboloid Geodesic, Wiedersehen Surface, Zoll Surface

Portions of this entry contributed by Todd Rowland

Explore with Wolfram|Alpha

References

Cipra, B. What's Happening in the Mathematical Sciences, Vol. 1. Providence, RI: Amer. Math. Soc., pp. 21-25, 1993.Tietze, H. Famous Problems of Mathematics: Solved and Unsolved Mathematics Problems from Antiquity to Modern Times. New York: Graylock Press, 1965.Weinstock, R. Calculus of Variations, with Applications to Physics and Engineering. New York: Dover, 1974.Weyl, H. Mathematische Analyse des Raumproblems: Was Ist Materie? Berlin: Wissenschaftl. Buchgesellschaft, 1923.

Referenced on Wolfram|Alpha

Geodesic

Cite this as:

Rowland, Todd and Weisstein, Eric W. "Geodesic." From MathWorld--A Wolfram Web Resource. https://mathworld.wolfram.com/Geodesic.html

Subject classifications