0% found this document useful (0 votes)
40 views

Multi Objective Optimization - MP

The document contains a random number table with 100 random numbers between 0 and 1 in 10 columns and 10 rows. It also contains examples of using a genetic algorithm to solve two optimization problems. In the first problem, the genetic algorithm is used to find the maximum of a function subject to bounds on its variables. In the second problem, the non-dominated sorting algorithm is applied to a multi-objective optimization problem to identify the Pareto fronts and calculate the crowding distance of chromosomes on the second highest ranked front.

Uploaded by

NEERAJ SINGH
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
40 views

Multi Objective Optimization - MP

The document contains a random number table with 100 random numbers between 0 and 1 in 10 columns and 10 rows. It also contains examples of using a genetic algorithm to solve two optimization problems. In the first problem, the genetic algorithm is used to find the maximum of a function subject to bounds on its variables. In the second problem, the non-dominated sorting algorithm is applied to a multi-objective optimization problem to identify the Pareto fronts and calculate the crowding distance of chromosomes on the second highest ranked front.

Uploaded by

NEERAJ SINGH
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Random number table

𝑖 1 2 3 4 5 6 7 8 9 10
𝑅𝑖 0.8261 0.7457 0.3100 0.7504 0.2752 0.6522 0.2129 0.7258 0.2978 0.9993
𝑖 11 12 13 14 15 16 17 18 19 20
𝑅𝑖 0.2894 0.6596 0.0318 0.8721 0.7231 0.6852 0.0460 0.9982 0.0800 0.336
𝑖 21 22 23 24 25 26 27 28 29 30
𝑅𝑖 0.5972 0.4297 0.7969 0.8281 0.7205 0.5463 0.3417 0.8762 0.6264 0.7356
𝑖 31 32 33 34 35 36 37 38 39 40
𝑅𝑖 0.1550 0.6390 0.8218 0.1082 0.8567 0.7590 0.0336 0.6858 0.1632 0.2472
𝑖 41 42 43 44 45 46 47 48 49 50
𝑅𝑖 0.6709 0.7878 0.5989 0.0938 0.3256 0.0517 0.9671 0.8258 0.4917 0.6447
𝑖 51 52 53 54 55 56 57 58 59 60
𝑅𝑖 0.2363 0.1609 0.0857 0.6007 0.1790 0.7531 0.4361 0.7471 0.1594 0.4248
𝑖 61 62 63 64 65 66 67 68 69 70
𝑅𝑖 0.3246 0.7017 0.4816 0.9433 0.1547 0.7930 0.5988 0.127 0.9004 0.7882
𝑖 71 72 73 74 75 76 77 78 79 80
𝑅𝑖 0.4896 0.7080 0.8588 0.0574 0.6370 0.2163 0.0071 0.0066 0.90593 0.86444

Problem1:
Max 𝒇(𝒙)=𝒙𝟐𝟏 + 𝒙𝟐 + 𝟏
s.t. 𝒙𝑳 = 𝟎 ≤ 𝒙 ≤ 𝟓 = 𝒙𝑼 .
Sol:
𝑭𝒊𝒕𝒏𝒆𝒔𝒔 𝒇𝒖𝒏𝒄𝒕𝒊𝒐𝒏 𝑭(𝒙) = 𝒇(𝒙) = 𝒙𝟐𝟏 + 𝒙𝟐 + 𝟏

(𝒙𝑼 𝑳
𝒊 − 𝒙𝒊 )
𝒓𝒆𝒂𝒍(𝒙𝒊 ) = 𝒙𝑳𝒊 + 𝒅𝒆𝒄𝒐𝒅𝒆𝒅 𝒗𝒂𝒍𝒖𝒆 𝒐𝒇 (𝒙𝒊 )
𝟐𝒍𝒔𝒕𝒓 − 𝟏
𝑆2 𝑆1 𝑆0 𝑆2 𝑆1 𝑆0
𝒙𝟏 𝒙𝟐
↔ ↔
𝒍𝒔𝒕𝒓 −𝟏

𝐃𝐞𝐜𝐨𝐝𝐞𝐝 𝐯𝐚𝐥𝐮𝐞 (𝒙𝒊 ) = ∑ 𝑺𝒋 𝟐𝒋 = 𝒔𝟎 𝟐𝟎 + 𝒔𝟏 𝟐𝟏 + 𝒔𝟐 𝟐𝟐


𝒋=𝟎
𝒆. 𝒈. 𝒇𝒐𝒓 𝒄𝒉𝒓𝒐𝒎𝒐𝒔𝒐𝒎𝒆 𝑪𝟏 , 𝒙𝟏 = 𝟎. 𝟐𝟎 + 𝟏. 𝟐𝟏 + 𝟏. 𝟐𝟐 = 𝟔 𝒂𝒏𝒅 𝒙𝟐 = 𝟏. 𝟐𝟎 + 𝟎. 𝟐𝟏 + 𝟏. 𝟐𝟐 = 𝟓
5−0 30
Real (𝑥1 ) = 0 + 6 = = 4.28
7 7

Initial population (𝑁𝑔𝑒𝑛 = 0) Mating pool or Gene pool


Chr. chromosome 𝒙 𝑭(𝒙) Chr. chromosome 𝒙 𝑭(𝒙)
No. No.
𝐶1 1 1 0 1 0 1 (4.68,5) 42 𝒔𝒆𝒍𝒆𝒄𝒕𝒊𝒐𝒏 𝑃1 0 1 1 1 0 1 (3,5) 15
𝐶2 0 1 0 1 0 1 (2,5) 10 → 𝑃2 0 1 0 1 0 1 (2,5) 10
𝐶3 0 1 1 1 0 1 (3,5) 15 𝑃3 0 1 1 1 0 1 (3,5) 15
𝐶4 0 0 1 0 1 1 (1,3) 5 𝑃4 1 1 0 1 0 1 (6,5) 42
1. Selection (Tournament selection):

(i) 𝑷𝒊𝒄𝒌 𝒂 𝒓𝒂𝒏𝒅𝒐𝒎 𝒏𝒖𝒎𝒃𝒆𝒓 𝑹𝒋


(ii) 𝒊𝒇 𝒌 − 𝟏 ≤ 𝑹𝒋 × 𝒏𝒄𝒉𝒓 ≤ 𝒌 𝒕𝒉𝒆𝒏 𝒔𝒆𝒍𝒆𝒄𝒕 𝑪𝑲 𝒄𝒉𝒓𝒐𝒎𝒐𝒔𝒐𝒎𝒆 ; 𝒌 = 𝟏, 𝟐, 𝟑, 𝟒
(iii) After selecting two chromosomes, keep one of them in mating pool with highest fitness
value (for maximization) / lowest fitness value (for minimization).
(iv) Repeat step (i) to (iii) until required number of chromosomes are selected in mating pool.

2. Crossover (crossover probability 𝑷𝑪𝒓𝒐𝒔𝒔 = 𝟎. 𝟗𝟖) single site crossover

𝑹𝟑𝟑 × 𝟒 ≈ 𝟑. 𝟐 ; 𝑷𝟒 𝑹𝟑𝟒 × 𝟒 ≈ 𝟎. 𝟒 ; 𝑷𝟏 ; 𝑹𝟑𝟕 × 𝟒 ≈ 𝟎. 𝟏𝟐 ; 𝑷𝟏 𝑹𝟑𝟖 × 𝟒 ≈ 𝟐. 𝟒 ; 𝑷𝟑 ;


𝑹𝟑𝟓 = 𝟎. 𝟖𝟓𝟔𝟕 < 𝑷𝒄𝒓𝒐𝒔𝒔 ; 𝒄𝒓𝒐𝒔𝒔𝒐𝒗𝒆𝒓 𝒑𝒆𝒓𝒇𝒐𝒓𝒎𝒆𝒅 𝑹𝟑𝟗 = 𝟎. 𝟏𝟔𝟑𝟐 < 𝑷𝒄𝒓𝒐𝒔𝒔 ; 𝒄𝒓𝒐𝒔𝒔𝒐𝒗𝒆𝒓 𝒑𝒆𝒓𝒇𝒐𝒓𝒎𝒆𝒅
𝑹𝟑𝟔 × (𝒍𝒄𝒉𝒓 − 𝟏) = 𝑹𝟑𝟔 × 𝟓 𝑹𝟒𝟎 × (𝒍𝒄𝒉𝒓 − 𝟏) = 𝑹𝟒𝟎 × 𝟓
≈ 𝟑. 𝟖; 𝟒𝒕𝒉 𝒊𝒏𝒕𝒆𝒓𝒏𝒂𝒍 𝒔𝒊𝒕𝒆 𝒂𝒔 𝒂 𝒄𝒓𝒐𝒔𝒔𝒐𝒗𝒆𝒓 𝒔𝒊𝒕𝒆 ≈ 𝟏. 𝟎𝟖; 𝟐𝒏𝒅 𝒊𝒏𝒕𝒆𝒓𝒏𝒂𝒍 𝒔𝒊𝒕𝒆 𝒂𝒔 𝒂 𝒄𝒓𝒐𝒔𝒔𝒐𝒗𝒆𝒓 𝒔𝒊𝒕𝒆
P4 1 1 0 1 0 1 P1 0 1 1 1 0 1

P1 0 1 1 1 0 1 P3 0 1 1 1 0 1
↓ 𝒄𝒓𝒐𝒔𝒔𝒐𝒗𝒆𝒓 ↓ 𝒄𝒓𝒐𝒔𝒔𝒐𝒗𝒆𝒓
D1 1 1 0 1 0 1 D3 0 1 1 1 0 1

D2 0 1 1 1 0 1 D4 0 1 1 1 0 1
.

Chromosome after
crossover
Chr. chromosome 𝒙 𝑭(𝒙)
No.
𝐷1 1 1 0 1 0 1 (6,5) 42
𝐷2 0 1 1 1 0 1 (3,5) 15
𝐷3 0 1 1 1 0 1 (3,5) 15
𝐷4 0 1 1 1 0 1 (3,5) 15

3. Mutation (mutation probability 𝑷𝒎𝒖𝒕 = 𝟎. 𝟐):

List of Random number from 𝑹𝟒𝟏 to 𝑹𝟔𝟒

0.6709 0.7878 0.5989 0.0938 0.3256 0.0517


0.9671 0.8258 0.4917 0.6447 0.2363 0.1609
0.0857 0.6007 0.1790 0.7531 0.4361 0.7471
0.1594 0.4248 0.3246 0.7017 0.4816 0.9433

Chromosome after Mutation


Chr. chromosome 𝒙 𝑭(𝒙)
No. = (𝒙𝟏 , 𝒙𝟐 )
𝐷1′ 1 1 0 0 0 0 (6,0) 37
𝐷2′ 0 1 1 1 0 0 (3,4) 14
𝐷3′ 1 1 0 1 0 1 (6,5) 42
𝐷4′ 1 1 1 1 0 1 (7,5) 55

This complete one iteration of genetic algorithm.

𝑵𝒈𝒆𝒏 = 𝟏

𝑹𝒑 = 𝟎. 𝟎𝟐𝟑𝟒, 𝑹𝒑+𝟏 = 𝟎. 𝟔𝟕𝟖𝟗

0 1
1, 𝒍𝒄𝒉𝒓 − 𝒍𝒋𝒈 =4
Problem2:

The fitness functions 𝑚𝑖𝑛𝐹1 and 𝑚𝑖𝑛𝐹2 for a multi-objective optimization problem corresponding to
9 chromosomes are given in the below table as:
Chromosome 𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7 𝐶8 𝐶9
No.
Min 𝐹1 2 3 3 4 4 5 5 5 6
Min 𝐹2 7.5 6 7.5 5 6.5 4.5 6 7 6.5
Draw the pareto optimal fronts and assign rank to the fronts. Also write (calculate) the crowding
distance of each chromosome at 2𝑛𝑑 highest ranked front.
Solution:

Min 𝑭𝟏 is 2 Max 𝑭𝟏 is 6

Min 𝑭𝟐 is 4.5 Max 𝑭𝟐 is 7.5

Chromosome No. 𝐶1 𝐶2 𝐶3 𝐶4 𝐶5 𝐶6 𝐶7 𝐶8 𝐶9
Min 𝐹1 2 3 3 4 4 5 5 5 6
Min 𝐹2 7.5 6 7.5 5 6.5 4.5 6 7 6.5

𝟐𝒏𝒅 𝒉𝒊𝒈𝒉𝒆𝒔𝒕 𝒓𝒂𝒏𝒌𝒆𝒅 𝒇𝒐𝒓𝒏𝒕 𝒊𝒔 𝒇𝒓𝒐𝒏𝒕 𝒘𝒊𝒕𝒉 𝒄𝒉𝒓𝒐𝒎𝒐𝒔𝒐𝒎𝒆𝒔 𝑪𝟑 , 𝑪𝟓 𝒂𝒏𝒅 𝑪𝟕 .

Since chromosomes 𝑪𝟑 and 𝑪𝟕 are boundary chromosomes, crowding distance of 𝑪𝟑 and 𝑪𝟕 is ∞.

Crowding distance of 𝑪𝟓 is calculated as:


𝟏
Crowding distance of 𝑪𝟓 = 𝟐 [|𝑭𝟏 (𝑪𝟑 ) − 𝑭𝟏 (𝑪𝟕 )|] + |𝑭𝟐 (𝑪𝟑 ) − 𝑭𝟐 (𝑪𝟕 )|

𝟏 𝟏 𝟑. 𝟓
= [|𝟑 − 𝟓| + |𝟕. 𝟓 − 𝟔|] = (𝟐 + 𝟏. 𝟓) = = 𝟏. 𝟕𝟓
𝟐 𝟐 𝟐

You might also like