Optimal Meshing

134
Optimal Meshing Don Sheehy INRIA Saclay, France soon: University of Connecticut

description

The word optimal is used in different ways in mesh generation.  It could mean that the output is in some sense, "the best mesh" or that the algorithm is, by some measure, "the best algorithm".  One might hope that the best algorithm also produces the best mesh, but maybe some tradeoffs are necessary.  In this talk, I will survey several different notions of optimality in mesh generation and explore the different tradeoffs between them. The bias will be towards Delaunay/Voronoi methods.

Transcript of Optimal Meshing

Page 1: Optimal Meshing

Optimal Meshing

Don SheehyINRIA Saclay, France

soon: University of Connecticut

Page 2: Optimal Meshing

Mesh Generation

Page 3: Optimal Meshing

Mesh Generation

bias: Delaunay/Voronoi Refinement

Page 4: Optimal Meshing

Mesh Generation

bias: Delaunay/Voronoi Refinementwhy: We want theoretical guarantees.

Page 5: Optimal Meshing

Mesh Generation

bias: Delaunay/Voronoi Refinementwhy: We want theoretical guarantees.

Everything will be in d dimensions, where d is a constant.

Page 6: Optimal Meshing

Mesh Generation

bias: Delaunay/Voronoi Refinementwhy: We want theoretical guarantees.

Everything will be in d dimensions, where d is a constant.

Constants that depend only on d will be hidden by (really) big-O’s.

Page 7: Optimal Meshing

Optimality

Page 8: Optimal Meshing

Optimality

Quality

Page 9: Optimal Meshing

Optimality

Quality Goal: Maximize element quality (many choices for what this means).

Page 10: Optimal Meshing

Optimality

Quality

Mesh Size

Goal: Maximize element quality (many choices for what this means).

Page 11: Optimal Meshing

Optimality

Quality

Mesh SizeGoal: Minimize Number of Vertices/Simplices

Goal: Maximize element quality (many choices for what this means).

Page 12: Optimal Meshing

Optimality

Quality

Mesh SizeGoal: Minimize Number of Vertices/Simplices

Also: Graded according to density/sizing function.

Goal: Maximize element quality (many choices for what this means).

Page 13: Optimal Meshing

Optimality

Quality

Mesh Size

Running Time

Goal: Minimize Number of Vertices/Simplices

Also: Graded according to density/sizing function.

Goal: Maximize element quality (many choices for what this means).

Page 14: Optimal Meshing

Optimality

Quality

Mesh Size

Running Time

Goal: Minimize Number of Vertices/Simplices

Also: Graded according to density/sizing function.

Goal: Maximize element quality (many choices for what this means).

Goal: O(n log n + m) time.

Page 15: Optimal Meshing

Optimality

Quality

Mesh Size

Running Time

Goal: Minimize Number of Vertices/Simplices

Also: Graded according to density/sizing function.

Goal: Maximize element quality (many choices for what this means).

Goal: O(n log n + m) time.

The emphasis will be on asymptotic bounds and minimum requirements so as to produce the most general lower bounds.

Page 16: Optimal Meshing

Quality

Page 17: Optimal Meshing

Quality

Many different/competing notions of quality.

Page 18: Optimal Meshing

Quality

Many different/competing notions of quality.

We will focus on those that yield theoretical guarantees.

Page 19: Optimal Meshing

Quality

Many different/competing notions of quality.

We will focus on those that yield theoretical guarantees.

This talk: Voronoi Aspect Ratio

Page 20: Optimal Meshing

Quality

Many different/competing notions of quality.

We will focus on those that yield theoretical guarantees.

This talk: Voronoi Aspect Ratio

rv

Rvv

Page 21: Optimal Meshing

Quality

Many different/competing notions of quality.

We will focus on those that yield theoretical guarantees.

This talk: Voronoi Aspect Ratio

Issues:

rv

Rvv

Page 22: Optimal Meshing

Quality

Many different/competing notions of quality.

We will focus on those that yield theoretical guarantees.

This talk: Voronoi Aspect Ratio

Issues:slivers

rv

Rvv

Page 23: Optimal Meshing

Quality

Many different/competing notions of quality.

We will focus on those that yield theoretical guarantees.

This talk: Voronoi Aspect Ratio

Issues:sliversgeometric stability rv

Rvv

Page 24: Optimal Meshing

Quality

Many different/competing notions of quality.

We will focus on those that yield theoretical guarantees.

This talk: Voronoi Aspect Ratio

Issues:sliversgeometric stabilitypost-processing/smoothing

rv

Rvv

Page 25: Optimal Meshing

Mesh Size

Page 26: Optimal Meshing

Mesh Size

The feature size function: fP (x) = min{r : |ball(x, r) \ P | � 2}

Page 27: Optimal Meshing

Mesh Size

The feature size measure: µP (⌦) =

Z

1

fP (x)ddx

The feature size function: fP (x) = min{r : |ball(x, r) \ P | � 2}

Page 28: Optimal Meshing

Mesh Size

The feature size measure: µP (⌦) =

Z

1

fP (x)ddx

|M | = ⇥(µP (⌦))

The feature size function: fP (x) = min{r : |ball(x, r) \ P | � 2}

Page 29: Optimal Meshing

Mesh Size

Assumes boundary has “small” complexity.

The feature size measure: µP (⌦) =

Z

1

fP (x)ddx

|M | = ⇥(µP (⌦))

The feature size function: fP (x) = min{r : |ball(x, r) \ P | � 2}

Page 30: Optimal Meshing

Mesh Size

Assumes boundary has “small” complexity.

The feature size measure: µP (⌦) =

Z

1

fP (x)ddx

|M | = ⇥(µP (⌦))

The feature size function: fP (x) = min{r : |ball(x, r) \ P | � 2}

Page 31: Optimal Meshing

Mesh Size

Page 32: Optimal Meshing

Mesh Size

vRv

rv

Page 33: Optimal Meshing

Mesh Size

vRv

rv

Page 34: Optimal Meshing

Mesh Size

vRv

rv

Page 35: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

Page 36: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 37: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 38: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

Z

Vor(v)

dx

fP (x)d

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 39: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

Z

Vor(v)

dx

r

dv

Z

Vor(v)

dx

fP (x)d

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 40: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

Z

Bv

dx

r

dv

Z

Vor(v)

dx

r

dv

Z

Vor(v)

dx

fP (x)d

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 41: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

= V✓Rv

rv

◆d

Z

Bv

dx

r

dv

Z

Vor(v)

dx

r

dv

Z

Vor(v)

dx

fP (x)d

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 42: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

= V✓Rv

rv

◆d

Z

Bv

dx

r

dv

Z

Vor(v)

dx

r

dv

Z

Vor(v)

dx

fP (x)d

Z

bv

dx

fP (x)d

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 43: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

Z

bv

dx

(KRv)d = V

✓Rv

rv

◆d

Z

Bv

dx

r

dv

Z

Vor(v)

dx

r

dv

Z

Vor(v)

dx

fP (x)d

Z

bv

dx

fP (x)d

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 44: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

V✓

rvKRv

◆d

=

Z

bv

dx

(KRv)d = V

✓Rv

rv

◆d

Z

Bv

dx

r

dv

Z

Vor(v)

dx

r

dv

Z

Vor(v)

dx

fP (x)d

Z

bv

dx

fP (x)d

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 45: Optimal Meshing

Mesh Size

vRv

rv8x 2 Vor(v) : rv fP (x) KRv

V✓

rvKRv

◆d

=

Z

bv

dx

(KRv)d = V

✓Rv

rv

◆d

Z

Bv

dx

r

dv

Z

Vor(v)

dx

r

dv

Z

Vor(v)

dx

fP (x)d

Z

bv

dx

fP (x)d

✓1

K⌧

◆d

µP (Vor(v)) ⌧d

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 46: Optimal Meshing

Mesh Size

vRv

rv

There is at most and at least some constant amount of mass

in each Voronoi cell!

8x 2 Vor(v) : rv fP (x) KRv

V✓

rvKRv

◆d

=

Z

bv

dx

(KRv)d = V

✓Rv

rv

◆d

Z

Bv

dx

r

dv

Z

Vor(v)

dx

r

dv

Z

Vor(v)

dx

fP (x)d

Z

bv

dx

fP (x)d

✓1

K⌧

◆d

µP (Vor(v)) ⌧d

Prove your algorithm achieves thisalgorithm specific (not for this talk)

Page 47: Optimal Meshing

Mesh Size

Page 48: Optimal Meshing

Mesh Size

Domain: ⌦ ⇢ Rd

Page 49: Optimal Meshing

Mesh SizeM�

= {v 2 M | Vor(v) ✓ ⌦}

Domain: ⌦ ⇢ Rd

Page 50: Optimal Meshing

Mesh SizeM�

= {v 2 M | Vor(v) ✓ ⌦} M+= {v 2 M | Vor(v) \ ⌦ 6= ;}

Domain: ⌦ ⇢ Rd

Page 51: Optimal Meshing

Mesh SizeM�

= {v 2 M | Vor(v) ✓ ⌦} M+= {v 2 M | Vor(v) \ ⌦ 6= ;}

Domain: ⌦ ⇢ Rd

µP (⌦)

Page 52: Optimal Meshing

Mesh SizeM�

= {v 2 M | Vor(v) ✓ ⌦} M+= {v 2 M | Vor(v) \ ⌦ 6= ;}

Domain: ⌦ ⇢ Rd

µP (⌦)

X

v2M

µP (Vor(v) \ ⌦)

Page 53: Optimal Meshing

Mesh SizeM�

= {v 2 M | Vor(v) ✓ ⌦} M+= {v 2 M | Vor(v) \ ⌦ 6= ;}

Domain: ⌦ ⇢ Rd

µP (⌦)

X

v2M

µP (Vor(v) \ ⌦)

=

Page 54: Optimal Meshing

Mesh SizeM�

= {v 2 M | Vor(v) ✓ ⌦} M+= {v 2 M | Vor(v) \ ⌦ 6= ;}

Domain: ⌦ ⇢ Rd

µP (⌦)

X

v2M

µP (Vor(v) \ ⌦)

=

|M+| ⌧d

Page 55: Optimal Meshing

Mesh SizeM�

= {v 2 M | Vor(v) ✓ ⌦} M+= {v 2 M | Vor(v) \ ⌦ 6= ;}

Domain: ⌦ ⇢ Rd

µP (⌦)

X

v2M

µP (Vor(v) \ ⌦)

=

|M+| ⌧d|M�|✓

1

K⌧

◆d

Page 56: Optimal Meshing

Mesh SizeM�

= {v 2 M | Vor(v) ✓ ⌦} M+= {v 2 M | Vor(v) \ ⌦ 6= ;}

Domain: ⌦ ⇢ Rd

µP (⌦)

X

v2M

µP (Vor(v) \ ⌦)

=

|M+| ⌧d

Bounds are tight when |M+| ⇡ |M�|

|M�|✓

1

K⌧

◆d

Page 57: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

Page 58: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

Order the points.

Page 59: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

Order the points.

Page 60: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

Order the points.

Page 61: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

Order the points.

Page 62: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

Order the points.

Page 63: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

Order the points.

Page 64: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

pi

Order the points.

Page 65: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

a = !pi "NN(pi)!

pi

Order the points.

Page 66: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

b = !pi " 2NN(pi)!

a = !pi "NN(pi)!

pi

Order the points.

Page 67: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

b = !pi " 2NN(pi)!

a = !pi "NN(pi)!

pi

The pacing of the ith point is !i =b

a.

Order the points.

Page 68: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

b = !pi " 2NN(pi)!

a = !pi "NN(pi)!

pi

The pacing of the ith point is !i =b

a.

Let ! be the geometric mean, so!

log !i = n log !.

Order the points.

Page 69: Optimal Meshing

Mesh SizeTight per-instance bounds on the mesh size can be expressed in terms of the “pacing”.

b = !pi " 2NN(pi)!

a = !pi "NN(pi)!

pi

The pacing of the ith point is !i =b

a.

Let ! be the geometric mean, so!

log !i = n log !.

! is the pacing of the ordering.

Order the points.

Page 70: Optimal Meshing

Mesh SizeWe can write the feature size measure as a telescoping sum.

Page 71: Optimal Meshing

Mesh SizeWe can write the feature size measure as a telescoping sum.

Pi = {p1, . . . , pi}

Page 72: Optimal Meshing

Mesh SizeWe can write the feature size measure as a telescoping sum.

Pi = {p1, . . . , pi}

µP = µP2+

n!

i=3

"

µPi! µPi!1

#

Page 73: Optimal Meshing

Mesh SizeWe can write the feature size measure as a telescoping sum.

Pi = {p1, . . . , pi}

effect of adding the ith point.

µP = µP2+

n!

i=3

"

µPi! µPi!1

#

Page 74: Optimal Meshing

Mesh SizeWe can write the feature size measure as a telescoping sum.

Pi = {p1, . . . , pi}

effect of adding the ith point.

µP = µP2+

n!

i=3

"

µPi! µPi!1

#

µPi(!)! µPi!1

(!) = "(1 + log !i)

When the boundary is “simple” and the first two points are not too close compared to the diameter,

Page 75: Optimal Meshing

Mesh SizeWe can write the feature size measure as a telescoping sum.

Pi = {p1, . . . , pi}

effect of adding the ith point.

µP = µP2+

n!

i=3

"

µPi! µPi!1

#

µPi(!)! µPi!1

(!) = "(1 + log !i)

When the boundary is “simple” and the first two points are not too close compared to the diameter,

Thus,

µP (⌦) = µP2(⌦) +⇥(n+ n log �)

Page 76: Optimal Meshing

Mesh SizeWe can write the feature size measure as a telescoping sum.

Pi = {p1, . . . , pi}

effect of adding the ith point.

µP = µP2+

n!

i=3

"

µPi! µPi!1

#

µPi(!)! µPi!1

(!) = "(1 + log !i)

When the boundary is “simple” and the first two points are not too close compared to the diameter,

Thus,

µP (⌦) = µP2(⌦) +⇥(n+ n log �)Measure induced by just two points.

Page 77: Optimal Meshing

Mesh SizeWe can write the feature size measure as a telescoping sum.

Pi = {p1, . . . , pi}

effect of adding the ith point.

µP = µP2+

n!

i=3

"

µPi! µPi!1

#

µPi(!)! µPi!1

(!) = "(1 + log !i)

When the boundary is “simple” and the first two points are not too close compared to the diameter,

Thus,

µP (⌦) = µP2(⌦) +⇥(n+ n log �)Measure induced by just two points.

Output Mesh Size

Page 78: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 79: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 80: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 81: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 82: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 83: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 84: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 85: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 86: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 87: Optimal Meshing

Mesh SizeThe previous bound implies that there is only one necessary but insufficient condition for the output size to be superlinear in the number of input points.

[ Picture of bad case ]

Page 88: Optimal Meshing

Running Time

Page 89: Optimal Meshing

Running Time

In an incremental construction, the points are added one at a time.

Page 90: Optimal Meshing

Running Time

In an incremental construction, the points are added one at a time.

Where is the work?

Page 91: Optimal Meshing

Running Time

In an incremental construction, the points are added one at a time.

Where is the work?

1. Point Location O(log n) per input vertex

Page 92: Optimal Meshing

Running Time

In an incremental construction, the points are added one at a time.

Where is the work?

1. Point Location

2. Local Updates

O(log n) per input vertex

O(1) per vertex

Page 93: Optimal Meshing

Running Time

In an incremental construction, the points are added one at a time.

Where is the work?

1. Point Location

2. Local Updates

O(log n) per input vertex

O(1) per vertex

Goal: O(n log n + m)

Page 94: Optimal Meshing

Running Time

Page 95: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

Page 96: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

2 Avoid the one bad case. Use hierarchical structure.

Page 97: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

2 Avoid the one bad case. Use hierarchical structure.

3 Preprocess the input vertices for fast point location.

Page 98: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

Page 99: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

Incremental Construction

Page 100: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

Incremental Construction

Recover input (vertices or features)

Page 101: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

Incremental Construction

Recover input (vertices or features)

Refine

Page 102: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

Incremental Construction

Recover input (vertices or features)

Refine an

Page 103: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

Incremental Construction

Recover input (vertices or features)

Refine

Loop

an

Page 104: Optimal Meshing

Running Time

1 Keep it quality. Keep it sparse.

Incremental Construction

Recover input (vertices or features)

Refine

Loop

an

Since the mesh is always quality, we avoid the worst case for Voronoi diagrams.

Insertions only require a constant number of local updates.

Page 105: Optimal Meshing

Running Time

2 Avoid the one bad case.

Page 106: Optimal Meshing

Running Time

2 Avoid the one bad case.

Page 107: Optimal Meshing

Running Time

2 Avoid the one bad case.

Page 108: Optimal Meshing

Running Time

2 Avoid the one bad case.

If you see a big empty annulus, do something different.

Page 109: Optimal Meshing

Running Time

2 Avoid the one bad case.

If you see a big empty annulus, do something different. - hierarchies

Page 110: Optimal Meshing

Running Time

2 Avoid the one bad case.

If you see a big empty annulus, do something different. - hierarchies - delayed input

Page 111: Optimal Meshing

Running Time

2 Avoid the one bad case.

If you see a big empty annulus, do something different. - hierarchies - delayed input

Linear-size meshes are possible by relaxing the quality condition for this one case. [MPS08, HMPS09, MPS11, S12, MSV13]

Page 112: Optimal Meshing

Running Time

3 Preprocess the input vertices for fast point location.

Page 113: Optimal Meshing

Running Time

3 Preprocess the input vertices for fast point location.

Page 114: Optimal Meshing

Running Time

3 Preprocess the input vertices for fast point location.

Page 115: Optimal Meshing

Running Time

3 Preprocess the input vertices for fast point location.

Page 116: Optimal Meshing

Running Time

3 Preprocess the input vertices for fast point location.

Page 117: Optimal Meshing

Running Time

3 Preprocess the input vertices for fast point location.

How many steps?

Page 118: Optimal Meshing

Running Time

3 Preprocess the input vertices for fast point location.

How many steps?If starting from nearest inserted input point, we only need to take a constant number of steps.

Page 119: Optimal Meshing

Running Time

3 Preprocess the input vertices for fast point location.

How many steps?If starting from nearest inserted input point, we only need to take a constant number of steps.

Ordering input points takes O(n log n) time.

Page 120: Optimal Meshing

Overview

Page 121: Optimal Meshing

Overview

A Defense of Theory:

Page 122: Optimal Meshing

Overview

A Defense of Theory:General lower bounds

Page 123: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Page 124: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:

Page 125: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.

Page 126: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.We focused on Voronoi Aspect Ratio

Page 127: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.We focused on Voronoi Aspect Ratio

Optimal Mesh Size:

Page 128: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.We focused on Voronoi Aspect Ratio

Optimal Mesh Size:The feature size measure determines mesh size.

Page 129: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.We focused on Voronoi Aspect Ratio

Optimal Mesh Size:The feature size measure determines mesh size.The pacing determines the feature size measure.

Page 130: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.We focused on Voronoi Aspect Ratio

Optimal Mesh Size:The feature size measure determines mesh size.The pacing determines the feature size measure.

Algorithmic suggestions for optimal Running time:

Page 131: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.We focused on Voronoi Aspect Ratio

Optimal Mesh Size:The feature size measure determines mesh size.The pacing determines the feature size measure.

Algorithmic suggestions for optimal Running time:Use the Sparse Meshing paradigm.

Page 132: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.We focused on Voronoi Aspect Ratio

Optimal Mesh Size:The feature size measure determines mesh size.The pacing determines the feature size measure.

Algorithmic suggestions for optimal Running time:Use the Sparse Meshing paradigm.Adapt to large pacing.

Page 133: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.We focused on Voronoi Aspect Ratio

Optimal Mesh Size:The feature size measure determines mesh size.The pacing determines the feature size measure.

Algorithmic suggestions for optimal Running time:Use the Sparse Meshing paradigm.Adapt to large pacing.Preprocess for walk-based point location

Page 134: Optimal Meshing

Overview

A Defense of Theory:General lower boundsTheory can guide practice

Mesh Quality:Many choices.We focused on Voronoi Aspect Ratio

Optimal Mesh Size:The feature size measure determines mesh size.The pacing determines the feature size measure.

Algorithmic suggestions for optimal Running time:Use the Sparse Meshing paradigm.Adapt to large pacing.Preprocess for walk-based point location

Thank You