Skip to main content
fixed broken link to springerlink.com; added full citations in tooltips
Source Link

Here are some results for a particular random model of lines.

The equation for any line in the plane may be written $p=x\cos\theta+y\sin\theta$, with $0\leq\theta<\pi$.

The Poisson line tessellation is a random process where one places lines randomly according to a Poisson point process with parameter $\tau/\pi$ in this $(p,\theta)$ strip and with uniformly distributed orientation. This generates a system of generic random lines almost surely. I'm not sure what the state of the art is, but according to this 1964 paper by R.E. Milesthis 1964 paper by R.E. Miles (see also its sequelits sequel which contains sketches of proofs), the first three moments of the distribution of areas of the polygons resulting from the intersections of these lines are $\pi/\tau^2$, $\pi^4/2\tau^4$ and $4\pi^7/7\tau^6$.

Some numerical results for the first seven moments of the area (with parameter 1 so that $\tau=\pi$ in the notation above) from generating 250,000 polygons are in a 2007 paper by Michel and Paroux2007 paper by Michel and Paroux: 0.31813 (exact is $1/\pi\approx0.31831$), 0.5009 (exact is $1/2$), 1.8042 (exact is $4\pi/7\approx1.7952$), 11.47, 113.9, 1610, 29400 (where I've rounded off some unnecessary digits from their Fig. 8). They also have a figure showing the cumulative distribution function for the areas:

Figure 13 from Michel and Paroux

The behavior near zero area is apparently explained by the following asymptotic result for the probability density function for the areas (part of Theorem 6 in Miles's paper above): $P(A\in dx)=c\tau A^{-1/2}+O(\tau^2)$, where $c=\frac{1}{3\pi}\int_0^\pi d\phi\int_0^{\pi-\phi}d\psi \sqrt{2\sin\phi\sin\psi\sin(\phi+\psi)}$ which Mathematica tells me is about 0.32311.

Here are some results for a particular random model of lines.

The equation for any line in the plane may be written $p=x\cos\theta+y\sin\theta$, with $0\leq\theta<\pi$.

The Poisson line tessellation is a random process where one places lines randomly according to a Poisson point process with parameter $\tau/\pi$ in this $(p,\theta)$ strip and with uniformly distributed orientation. This generates a system of generic random lines almost surely. I'm not sure what the state of the art is, but according to this 1964 paper by R.E. Miles (see also its sequel which contains sketches of proofs), the first three moments of the distribution of areas of the polygons resulting from the intersections of these lines are $\pi/\tau^2$, $\pi^4/2\tau^4$ and $4\pi^7/7\tau^6$.

Some numerical results for the first seven moments of the area (with parameter 1 so that $\tau=\pi$ in the notation above) from generating 250,000 polygons are in a 2007 paper by Michel and Paroux: 0.31813 (exact is $1/\pi\approx0.31831$), 0.5009 (exact is $1/2$), 1.8042 (exact is $4\pi/7\approx1.7952$), 11.47, 113.9, 1610, 29400 (where I've rounded off some unnecessary digits from their Fig. 8). They also have a figure showing the cumulative distribution function for the areas:

Figure 13 from Michel and Paroux

The behavior near zero area is apparently explained by the following asymptotic result for the probability density function for the areas (part of Theorem 6 in Miles's paper above): $P(A\in dx)=c\tau A^{-1/2}+O(\tau^2)$, where $c=\frac{1}{3\pi}\int_0^\pi d\phi\int_0^{\pi-\phi}d\psi \sqrt{2\sin\phi\sin\psi\sin(\phi+\psi)}$ which Mathematica tells me is about 0.32311.

Here are some results for a particular random model of lines.

The equation for any line in the plane may be written $p=x\cos\theta+y\sin\theta$, with $0\leq\theta<\pi$.

The Poisson line tessellation is a random process where one places lines randomly according to a Poisson point process with parameter $\tau/\pi$ in this $(p,\theta)$ strip and with uniformly distributed orientation. This generates a system of generic random lines almost surely. I'm not sure what the state of the art is, but according to this 1964 paper by R.E. Miles (see also its sequel which contains sketches of proofs), the first three moments of the distribution of areas of the polygons resulting from the intersections of these lines are $\pi/\tau^2$, $\pi^4/2\tau^4$ and $4\pi^7/7\tau^6$.

Some numerical results for the first seven moments of the area (with parameter 1 so that $\tau=\pi$ in the notation above) from generating 250,000 polygons are in a 2007 paper by Michel and Paroux: 0.31813 (exact is $1/\pi\approx0.31831$), 0.5009 (exact is $1/2$), 1.8042 (exact is $4\pi/7\approx1.7952$), 11.47, 113.9, 1610, 29400 (where I've rounded off some unnecessary digits from their Fig. 8). They also have a figure showing the cumulative distribution function for the areas:

Figure 13 from Michel and Paroux

The behavior near zero area is apparently explained by the following asymptotic result for the probability density function for the areas (part of Theorem 6 in Miles's paper above): $P(A\in dx)=c\tau A^{-1/2}+O(\tau^2)$, where $c=\frac{1}{3\pi}\int_0^\pi d\phi\int_0^{\pi-\phi}d\psi \sqrt{2\sin\phi\sin\psi\sin(\phi+\psi)}$ which Mathematica tells me is about 0.32311.

fix image
Source Link
j.c.
  • 13.6k
  • 3
  • 52
  • 90

Here are some results for a particular random model of lines.

The equation for any line in the plane may be written $p=x\cos\theta+y\sin\theta$, with $0\leq\theta<\pi$.

The Poisson line tessellation is a random process where one places lines randomly according to a Poisson point process with parameter $\tau/\pi$ in this $(p,\theta)$ strip and with uniformly distributed orientation. This generates a system of generic random lines almost surely. I'm not sure what the state of the art is, but according to this 1964 paper by R.E. Miles (see also its sequel which contains sketches of proofs), the first three moments of the distribution of areas of the polygons resulting from the intersections of these lines are $\pi/\tau^2$, $\pi^4/2\tau^4$ and $4\pi^7/7\tau^6$.

Some numerical results for the first seven moments of the area (with parameter 1 so that $\tau=\pi$ in the notation above) from generating 250,000 polygons are in a 2007 paper by Michel and Paroux: 0.31813 (exact is $1/\pi\approx0.31831$), 0.5009 (exact is $1/2$), 1.8042 (exact is $4\pi/7\approx1.7952$), 11.47, 113.9, 1610, 29400 (where I've rounded off some unnecessary digits from their Fig. 8). They also have a figure showing the cumulative distribution function for the areas:

Figure 13 from Michel and ParouxFigure 13 from Michel and Paroux

The behavior near zero area is apparently explained by the following asymptotic result for the probability density function for the areas (part of Theorem 6 in Miles's paper above): $P(A\in dx)=c\tau A^{-1/2}+O(\tau^2)$, where $c=\frac{1}{3\pi}\int_0^\pi d\phi\int_0^{\pi-\phi}d\psi \sqrt{2\sin\phi\sin\psi\sin(\phi+\psi)}$ which Mathematica tells me is about 0.32311.

Here are some results for a particular random model of lines.

The equation for any line in the plane may be written $p=x\cos\theta+y\sin\theta$, with $0\leq\theta<\pi$.

The Poisson line tessellation is a random process where one places lines randomly according to a Poisson point process with parameter $\tau/\pi$ in this $(p,\theta)$ strip and with uniformly distributed orientation. This generates a system of generic random lines almost surely. I'm not sure what the state of the art is, but according to this 1964 paper by R.E. Miles (see also its sequel which contains sketches of proofs), the first three moments of the distribution of areas of the polygons resulting from the intersections of these lines are $\pi/\tau^2$, $\pi^4/2\tau^4$ and $4\pi^7/7\tau^6$.

Some numerical results for the first seven moments of the area (with parameter 1 so that $\tau=\pi$ in the notation above) from generating 250,000 polygons are in a 2007 paper by Michel and Paroux: 0.31813 (exact is $1/\pi\approx0.31831$), 0.5009 (exact is $1/2$), 1.8042 (exact is $4\pi/7\approx1.7952$), 11.47, 113.9, 1610, 29400 (where I've rounded off some unnecessary digits from their Fig. 8). They also have a figure showing the cumulative distribution function for the areas:

Figure 13 from Michel and Paroux

The behavior near zero area is apparently explained by the following asymptotic result for the probability density function for the areas (part of Theorem 6 in Miles's paper above): $P(A\in dx)=c\tau A^{-1/2}+O(\tau^2)$, where $c=\frac{1}{3\pi}\int_0^\pi d\phi\int_0^{\pi-\phi}d\psi \sqrt{2\sin\phi\sin\psi\sin(\phi+\psi)}$ which Mathematica tells me is about 0.32311.

Here are some results for a particular random model of lines.

The equation for any line in the plane may be written $p=x\cos\theta+y\sin\theta$, with $0\leq\theta<\pi$.

The Poisson line tessellation is a random process where one places lines randomly according to a Poisson point process with parameter $\tau/\pi$ in this $(p,\theta)$ strip and with uniformly distributed orientation. This generates a system of generic random lines almost surely. I'm not sure what the state of the art is, but according to this 1964 paper by R.E. Miles (see also its sequel which contains sketches of proofs), the first three moments of the distribution of areas of the polygons resulting from the intersections of these lines are $\pi/\tau^2$, $\pi^4/2\tau^4$ and $4\pi^7/7\tau^6$.

Some numerical results for the first seven moments of the area (with parameter 1 so that $\tau=\pi$ in the notation above) from generating 250,000 polygons are in a 2007 paper by Michel and Paroux: 0.31813 (exact is $1/\pi\approx0.31831$), 0.5009 (exact is $1/2$), 1.8042 (exact is $4\pi/7\approx1.7952$), 11.47, 113.9, 1610, 29400 (where I've rounded off some unnecessary digits from their Fig. 8). They also have a figure showing the cumulative distribution function for the areas:

Figure 13 from Michel and Paroux

The behavior near zero area is apparently explained by the following asymptotic result for the probability density function for the areas (part of Theorem 6 in Miles's paper above): $P(A\in dx)=c\tau A^{-1/2}+O(\tau^2)$, where $c=\frac{1}{3\pi}\int_0^\pi d\phi\int_0^{\pi-\phi}d\psi \sqrt{2\sin\phi\sin\psi\sin(\phi+\psi)}$ which Mathematica tells me is about 0.32311.

replaced http://i583.photobucket.com/ with https://i583.photobucket.com/
Source Link

Here are some results for a particular random model of lines.

The equation for any line in the plane may be written $p=x\cos\theta+y\sin\theta$, with $0\leq\theta<\pi$.

The Poisson line tessellation is a random process where one places lines randomly according to a Poisson point process with parameter $\tau/\pi$ in this $(p,\theta)$ strip and with uniformly distributed orientation. This generates a system of generic random lines almost surely. I'm not sure what the state of the art is, but according to this 1964 paper by R.E. Miles (see also its sequel which contains sketches of proofs), the first three moments of the distribution of areas of the polygons resulting from the intersections of these lines are $\pi/\tau^2$, $\pi^4/2\tau^4$ and $4\pi^7/7\tau^6$.

Some numerical results for the first seven moments of the area (with parameter 1 so that $\tau=\pi$ in the notation above) from generating 250,000 polygons are in a 2007 paper by Michel and Paroux: 0.31813 (exact is $1/\pi\approx0.31831$), 0.5009 (exact is $1/2$), 1.8042 (exact is $4\pi/7\approx1.7952$), 11.47, 113.9, 1610, 29400 (where I've rounded off some unnecessary digits from their Fig. 8). They also have a figure showing the cumulative distribution function for the areas:

Figure 13 from Michel and Paroux http://i583.photobucket.com/albums/ss275/jaspercrowne/michelperouxfig13.pngFigure 13 from Michel and Paroux

The behavior near zero area is apparently explained by the following asymptotic result for the probability density function for the areas (part of Theorem 6 in Miles's paper above): $P(A\in dx)=c\tau A^{-1/2}+O(\tau^2)$, where $c=\frac{1}{3\pi}\int_0^\pi d\phi\int_0^{\pi-\phi}d\psi \sqrt{2\sin\phi\sin\psi\sin(\phi+\psi)}$ which Mathematica tells me is about 0.32311.

Here are some results for a particular random model of lines.

The equation for any line in the plane may be written $p=x\cos\theta+y\sin\theta$, with $0\leq\theta<\pi$.

The Poisson line tessellation is a random process where one places lines randomly according to a Poisson point process with parameter $\tau/\pi$ in this $(p,\theta)$ strip and with uniformly distributed orientation. This generates a system of generic random lines almost surely. I'm not sure what the state of the art is, but according to this 1964 paper by R.E. Miles (see also its sequel which contains sketches of proofs), the first three moments of the distribution of areas of the polygons resulting from the intersections of these lines are $\pi/\tau^2$, $\pi^4/2\tau^4$ and $4\pi^7/7\tau^6$.

Some numerical results for the first seven moments of the area (with parameter 1 so that $\tau=\pi$ in the notation above) from generating 250,000 polygons are in a 2007 paper by Michel and Paroux: 0.31813 (exact is $1/\pi\approx0.31831$), 0.5009 (exact is $1/2$), 1.8042 (exact is $4\pi/7\approx1.7952$), 11.47, 113.9, 1610, 29400 (where I've rounded off some unnecessary digits from their Fig. 8). They also have a figure showing the cumulative distribution function for the areas:

Figure 13 from Michel and Paroux http://i583.photobucket.com/albums/ss275/jaspercrowne/michelperouxfig13.png

The behavior near zero area is apparently explained by the following asymptotic result for the probability density function for the areas (part of Theorem 6 in Miles's paper above): $P(A\in dx)=c\tau A^{-1/2}+O(\tau^2)$, where $c=\frac{1}{3\pi}\int_0^\pi d\phi\int_0^{\pi-\phi}d\psi \sqrt{2\sin\phi\sin\psi\sin(\phi+\psi)}$ which Mathematica tells me is about 0.32311.

Here are some results for a particular random model of lines.

The equation for any line in the plane may be written $p=x\cos\theta+y\sin\theta$, with $0\leq\theta<\pi$.

The Poisson line tessellation is a random process where one places lines randomly according to a Poisson point process with parameter $\tau/\pi$ in this $(p,\theta)$ strip and with uniformly distributed orientation. This generates a system of generic random lines almost surely. I'm not sure what the state of the art is, but according to this 1964 paper by R.E. Miles (see also its sequel which contains sketches of proofs), the first three moments of the distribution of areas of the polygons resulting from the intersections of these lines are $\pi/\tau^2$, $\pi^4/2\tau^4$ and $4\pi^7/7\tau^6$.

Some numerical results for the first seven moments of the area (with parameter 1 so that $\tau=\pi$ in the notation above) from generating 250,000 polygons are in a 2007 paper by Michel and Paroux: 0.31813 (exact is $1/\pi\approx0.31831$), 0.5009 (exact is $1/2$), 1.8042 (exact is $4\pi/7\approx1.7952$), 11.47, 113.9, 1610, 29400 (where I've rounded off some unnecessary digits from their Fig. 8). They also have a figure showing the cumulative distribution function for the areas:

Figure 13 from Michel and Paroux

The behavior near zero area is apparently explained by the following asymptotic result for the probability density function for the areas (part of Theorem 6 in Miles's paper above): $P(A\in dx)=c\tau A^{-1/2}+O(\tau^2)$, where $c=\frac{1}{3\pi}\int_0^\pi d\phi\int_0^{\pi-\phi}d\psi \sqrt{2\sin\phi\sin\psi\sin(\phi+\psi)}$ which Mathematica tells me is about 0.32311.

some more details
Source Link
j.c.
  • 13.6k
  • 3
  • 52
  • 90
Loading
Source Link
j.c.
  • 13.6k
  • 3
  • 52
  • 90
Loading