a new analysis of the lebmeasure algorithm for calculating hypervolume lyndon while walking fish...

25
A New Analysis of A New Analysis of the LebMeasure Algorithm the LebMeasure Algorithm for Calculating Hypervolume for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering The University of Western Australia

Upload: santos-mull

Post on 01-Apr-2015

224 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis ofA New Analysis ofthe LebMeasure Algorithmthe LebMeasure Algorithm

for Calculating Hypervolumefor Calculating Hypervolume

Lyndon While

Walking Fish GroupSchool of Computer Science & Software Engineering

The University of Western Australia

Page 2: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 2 of 25

Overview

Metrics for MOEAs Hypervolume LebMeasure and its behaviour Empirical data on the performance of LebMeasure A lower-bound on the complexity of LebMeasure The general case Conclusions and future work

Page 3: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 3 of 25

Metrics for MOEAs

A MOEA produces a front of mutually non-dominating solutions to a given problem

m points in n objectives To compare the performance of MOEAs,

we need metrics to compare fronts Many metrics have been proposed, of several types

cardinality-based metrics convergence-based metrics spread-based metrics volume-based metrics

Page 4: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 4 of 25

Hypervolume (S-metric, Lebesgue measure)

The hypervolume of a front is the size of the portion of objective space collectively dominated by the points on the front

Hypervolume captures in one scalar both the convergence and the spread of the front

Hypervolume has nicer mathematical properties than many other metrics

Hypervolume can be sensitive to scaling of objectivesand to extremal values

Hypervolume is expensive to calculate enter LebMeasure

Page 5: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 5 of 25

LebMeasure (LM)

Given a mutually non-dominating front S, LM calculates the hypervolume dominated exclusively

by the first point p, then discards p and processes the rest of S

If the hypervolume dominated exclusively by p is not “hyper-cuboid”, LM

lops off a hyper-cuboid that isdominated exclusively by p, and

replaces p with up to n “spawns” that collectively dominate the remainder of p’s exclusive hypervolume

A spawn is discarded immediately if it dominates no exclusive hypervolume, either because

it has a “zero” objective, or it is dominated by an unprocessed point

Page 6: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 6 of 25

LebMeasure in action

A dominates exclusively the yellow shape A lops off the pink hyper-cuboid A has three potential spawns:

A1 = (4,9,4)A2 = (6,7,4)A3 = (6,9,3)

But A2 is dominated by B, so it is discarded immediately

Page 7: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 7 of 25

A boost for LebMeasure

Some “spawns of spawns” are guaranteed to be dominated, so LM doesn’t need to generate them at all

This limits the maximum depth of the stack to m + n – 1

(6, 9, 4)

(9, 7, 5)

(1,12, 3)

(4, 2, 9)

(4, 9, 4)

(6, 9, 3)

(9, 7, 5)

(1,12, 3)

(4, 2, 9)

(1, 9, 4)

(4, 7, 4)

(4, 9, 3)

(6, 9, 3)

(9, 7, 5)

(1,12, 3)

(4, 2, 9)

A13

A12

A11

guaranteed tobe dominated}

A3

B

C

DD D

C C

BB

A A3

A1

Page 8: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 8 of 25

But…

This boost greatly reduces the space complexity of LM the maximum depth of the stack is linear in both m and n

But it does far less for the time complexity of LM note that the time complexity depends

not only on the number of stack slots used, but also on how many times each slot is used

We shall measure the time complexity of LM in terms of the number of points (and spawns, and spawns of spawns, etc) that actually contribute to the hypervolume

i.e. the number of hyper-cuboids that must be summed

Page 9: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 9 of 25

Running LebMeasure

1 5 ••• 5

2 4 ••• 4

3 3 ••• 3

4 2 ••• 2

5 1 ••• 1

n m = 2 m = 5 m = 8 m = 10

2 2 5 8 10

3 4 25 64 100

4 8 125 512 1,000

5 16 625 4,096 10,000

6 32 3,125 32,768 100,000

7 64 15,625 262,144 1,000,000

8 128 78,125 2,097,152 10,000,000

9 256 390,62516,777,21

6100,000,000

No. of hyper-cuboids = mn−1

m points in n objectives

Page 10: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 10 of 25

Running LebMeasure (in reverse order)

1 5 ••• 5

2 4 ••• 4

3 3 ••• 3

4 2 ••• 2

5 1 ••• 1

No. of hyper-cuboids = m

n m = 2 m = 5 m = 8 m = 10

2 2 5 8 10

3 2 5 8 10

4 2 5 8 10

5 2 5 8 10

6 2 5 8 10

7 2 5 8 10

8 2 5 8 10

9 2 5 8 10

m points in n objectives

Page 11: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 11 of 25

Running LebMeasure (in optimal order)

1 1 2 3 4 5 1 ••• 5

2 2 3 4 5 1 2 ••• 4

3 3 4 5 1 2 3 ••• 3

4 4 5 1 2 3 4 ••• 2

5 5 1 2 3 4 5 ••• 1

n m = 2 m = 3 m = 4 m = 5

2 2 3 4 5

3 2 3 4 5

4 4 6 8 10

5 4 17 23 29

6 8 17 88 112

7 8 35 88 549

8 16 105 180 549

9 16 105 558 1,115

10 32 213 2,248 3,421

11 32 641 2,248 14,083

12 64 641 4,528 70,899

13 64 1,289 13,708 70,889

14 128 3,873 54,976 142,309

15 128 3,873 54,976 428,449

16 256 7,761 110,160 1,721,605

17 256 23,297 331,128 8,618,577

No. of hyper-cuboids m(m!)((n−2)div m)((n – 2)mod m)!

m points in n objectives

Page 12: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 12 of 25

Running LebMeasure (first point only)

1 5 ••• 5

2 4 ••• 4

3 3 ••• 3

4 2 ••• 2

5 1 ••• 1

n m = 2 m = 5 m = 8 m = 10

2 1 1 1 1

3 3 9 15 19

4 7 61 169 271

5 15 369 1,695 3,439

6 31 2,101 15,961 40,951

7 63 11,529 144,495 468,559

8 127 61,741 1,273,609 5,217,031

9 255 325,089 11,012,415 56,953,279

No. of hyper-cuboids = mn−1 – (m – 1)n−1, i.e. O(mn−2)

m points in n objectives

Page 13: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 13 of 25

A lower-bound on the complexity of LebMeasure

We can determine a lower-bound on the worst-case complexity of LM by considering a single example

We will derive a recurrence for the number of hyper-cuboids summed for this example, then prove that the recurrence equals 2n−1

1 2 2 2 ••• 2

2 1 1 1 ••• 1

Page 14: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 14 of 25

The simple picture

12222

11222 12122 12212 12221

11112

11221 122111212111212 1211211122

11211 1211111121

Page 15: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 15 of 25

The recursive picture

12222

11212 12112

11112

11122

11222 12122 12212

1211111211

12211

11121

11221 12121

12221

Page 16: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 16 of 25

A recurrence

h(n,k) gives the number of hyper-cuboids summed for a point (or spawn) with n 2s, of which we can reduce k and still generate points that aren’t dominated by their relatives

hcs(n) gives the total number of hyper-cuboids summed for the example, with n objectives

1

0),1(1),(

1),1(k

iinhknh

kh

1)1,1()( nnhnhcs

Page 17: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 17 of 25

The recurrence in action

[h(4,4)]

(1,2,2,2,2)

(1,2,2,1,2)

[h(3,2)]

(1,2,2,2,1)

[h(3,3)]

(1,2,1,2,2)

[h(3,1)]

(1,1,2,2,2)

[h(3,0)]

Page 18: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 18 of 25

The recurrence solved

Simple expansion shows that

The paper gives a formal proof using mathematical induction

1

1

2)(

12),(

n

k

nhcs

kkh

Page 19: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 19 of 25

The general case

It is difficult to be certain what patterns of points will perform worst for LM

We will describe the behaviour of an illegal “beyond worst case” pattern

Illegal because some points dominate others

m m ••• m

m−1 m−1 ••• m−1

•••

•••

•••

1 1 ••• 1

Page 20: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 20 of 25

m points in 2 objectives

xi denotes the ith best value in objective x

Each vertical list has length m Total size m2

u1v1

Page 21: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 21 of 25

m points in 3 objectives

Each vertical list has length m Each 2-way sub-tree has size m2

Total size m3

u1v1w1

Page 22: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 22 of 25

m points in 4 objectives

denotes a k-way sub-tree

Each k-way sub-tree has size mk

Total size m4

k

12

3

12

3

12

3

12

3

u1v1w1x1

Page 23: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 23 of 25

A recurrence and its solution

Again, we can capture this behaviour as a recurrence

By simple expansion (and proved formally in the paper)

1

0

1

1

),,(

),,(11

10

m

i

y

i

iniqp(m,n)

zizq,y,z)q(xq(x,y,z)

,y,z)q(

m

i

ninmp1

),(

Page 24: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 24 of 25

Conclusions

LM is exponential in the number of objectives, in the worst case

Re-ordering the points often makes LM go faster,but the worst case is still exponential

the proof technique used for the “simple” case will also work for the “unreorderable” case

Page 25: A New Analysis of the LebMeasure Algorithm for Calculating Hypervolume Lyndon While Walking Fish Group School of Computer Science & Software Engineering

A New Analysis of LebMeasure Page 25 of 25

Future work

Try to make LM faster re-order the points re-order the objectives

Develop and refine other algorithms (e.g. HSO) possibly develop a hybrid algorithm

Prove that no polynomial-time algorithm existsfor calculating hypervolume