KEMBAR78
Machine Learning Lecture 5 (Genetic Algorithm) | PDF
Uploaded byssuser9cf5ab
12 views

Machine Learning Lecture 5 (Genetic Algorithm)

Machine Learning Lecture 5 containing Genetic Algorithms

Download to read offline
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Learning Machine Learning :: Session 3 Partha Dey, PhD Dept. of Mechanical Engineering Academy of Technology 18th November 2018 1 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Overview 1 Ant Colony Optimization 2 Genetic Algorithm Genetic Operators 3 The ACO/GA flowchart 4 The ACO/GA code 5 Acknowledgements 2 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route 3 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route 3 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route 3 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route 3 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route She first 1 finds out the inter-city distances — 3 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route She first 1 finds out the inter-city distances — Kol Ben Guw Che Mum Kol Ben 1560 Guw 524 2082 Che 1356 290 1884 Mum 1654 846 1034 2088 Source : www.distancecalculator.net 3 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route She first 1 finds out the inter-city distances — Kol Ben Guw Che Mum Kol Ben 1560 Guw 524 2082 Che 1356 290 1884 Mum 1654 846 1034 2088 Source : www.distancecalculator.net . . . then 2 lists down all the 4 ! ways to plan her tour, 3 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route She first 1 finds out the inter-city distances — Kol Ben Guw Che Mum Kol Ben 1560 Guw 524 2082 Che 1356 290 1884 Mum 1654 846 1034 2088 Source : www.distancecalculator.net . . . then 2 lists down all the 4 ! ways to plan her tour, and finally 3 picks the shortest of them. 3 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. BOM, 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. BOM, BLR, 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. BOM, BLR, MAA, 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. BOM, BLR, MAA, GAU 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . weaker untraversed trails . . . . . . 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . weaker untraversed trails . . . . . . evaporate fast 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . showing the way to an optimized path. . . 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . showing the way to an optimized path. . . for our actual tranveller to follow 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . showing the way to an optimized path. . . for our actual tranveller to follow, without doing sort(4 ! sums) 4 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) 5 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3, 11) 5 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (4, 6) 5 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3.6, 7.5) 5 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3.6, 7.5) Hence, we minimize A = 2πrh + 2πr2 subject to V = πr2 h ≄ 300 5 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3.6, 7.5) Hence, we minimize A = 2πrh + 2πr2 subject to V = πr2 h ≄ 300 We ‘encode’ the two dimensions (r and h) in the form of a gene 5 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3.6, 7.5) Hence, we minimize A = 2πrh + 2πr2 subject to V = πr2 h ≄ 300 We ‘encode’ the two dimensions (r and h) in the form of a gene 1 0 0 1 | 1 0 1 0 # Hackers : Any principle for encoding ? 5 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem (solved with GA) Let’s take a possible range of values of r and h : r (coded) | h (coded) 2.5 7→ 0 0 0 0 | 5.0 7→ 0 0 0 0 2.6 7→ 0 0 0 1 | 5.5 7→ 0 0 0 1 ... 7→ ... ... | ... 7→ ... ... 4.0 7→ 1 1 1 1 | 12.5 7→ 1 1 1 1 So, any can dimension, say (3.6, 7.5) could be encoded as : 6 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem (solved with GA) Let’s take a possible range of values of r and h : r (coded) | h (coded) 2.5 7→ 0 0 0 0 | 5.0 7→ 0 0 0 0 2.6 7→ 0 0 0 1 | 5.5 7→ 0 0 0 1 ... 7→ ... ... | ... 7→ ... ... 4.0 7→ 1 1 1 1 | 12.5 7→ 1 1 1 1 So, any can dimension, say (3.6, 7.5) could be encoded as : 1 0 1 1 | 0 1 0 1 6 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem (solved with GA) Let’s take a possible range of values of r and h : r (coded) | h (coded) 2.5 7→ 0 0 0 0 | 5.0 7→ 0 0 0 0 2.6 7→ 0 0 0 1 | 5.5 7→ 0 0 0 1 ... 7→ ... ... | ... 7→ ... ... 4.0 7→ 1 1 1 1 | 12.5 7→ 1 1 1 1 or, say (4, 6) would be encoded as : 1 1 1 1 | 0 0 1 0 6 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 1 1 1 1 0 0 1 0 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 We choose a crossover point, 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. → (3.6, 6) → (4, 7.5) 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled (to avoid stagnation of solutions) 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled (to avoid stagnation of solutions) These operators are called Crossover and Mutation. They mimic the natural evolution phenomena. 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled (to avoid stagnation of solutions) In each iteration (or generation), better performing solutions are selected with higher probability. So offsprings improve. 7 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The algorithm for ACO 8 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The algorithm for GA 8 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The R code for GA CAN - ga(type = real-valued, fitness = function(can.dim) GA.obj(can.dim), lower = c(2.5, 5), upper = c(4, 12.5), popSize = 20, maxiter = 10) GA.obj - function(rh) { p - 3.1416; r - rh[1]; h - rh[2] V - p*r*r*h; A - 2*p*r*h + 2*(p*r*r) if (r ... | r ...) return(-100) if (h ... | h ...) return(-200) if (V 300) return(-300) return(1000 - A) } 9 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements 10 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee 10 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya 10 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil 10 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity 10 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana 10 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana Prof. Ashoke Banerjee 10 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana Prof. Ashoke Banerjee Prof. Shubhabrata Dutta 10 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana Prof. Ashoke Banerjee Prof. Shubhabrata Dutta Prof. Kalyanmoy Deb and Prof. Marco Dorigo 10 / 10
Learning ML Partha Dey, PhD Ant Colony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana Prof. Ashoke Banerjee Prof. Shubhabrata Dutta Prof. Kalyanmoy Deb and Prof. Marco Dorigo All of you with your tried-and-tested patience 10 / 10

More Related Content

PDF
With saloni in ijarcsse
bysatish rana
 
PPT
Solving travelling salesman problem using firefly algorithm
byishmecse13
 
PDF
K046036367
byIJERA Editor
 
PDF
Comparison Study of Multiple Traveling Salesmen Problem using Genetic Algorithm
byIOSR Journals
 
PDF
An Implementational approach to genetic algorithms for TSP
bySougata Das
 
PDF
Traveling Salesman Problem Federico Greco
byzotteltambik
 
PPTX
Muzammil Adulrahman ppt on travelling salesman Problem Based On Mutation Gene...
byPetroleum Training Institute
 
PPTX
Travelling and salesman problem using ant colony optimization
byAkash Sethiya
 
With saloni in ijarcsse
bysatish rana
 
Solving travelling salesman problem using firefly algorithm
byishmecse13
 
K046036367
byIJERA Editor
 
Comparison Study of Multiple Traveling Salesmen Problem using Genetic Algorithm
byIOSR Journals
 
An Implementational approach to genetic algorithms for TSP
bySougata Das
 
Traveling Salesman Problem Federico Greco
byzotteltambik
 
Muzammil Adulrahman ppt on travelling salesman Problem Based On Mutation Gene...
byPetroleum Training Institute
 
Travelling and salesman problem using ant colony optimization
byAkash Sethiya
 

Similar to Machine Learning Lecture 5 (Genetic Algorithm)

PPTX
Travelling salesperson problem
byAkshayIngole17
 
PPTX
Travelling Salesman Problem
byShikha Gupta
 
PPTX
Operation research model for solving TSP
byDrGovindshaysharma
 
PDF
Traveling salesman problem__theory_and_applications
bySachin Kheveria
 
PDF
Traveling sales man problem research.pdf
bymoulayalih0101
 
PPTX
Travelling salesman problem using genetic algorithms
byShivank Shah
 
PPTX
Solving TSP using Hybrid GACO
byShruti Gandhi
 
PDF
A Survey of Solving Travelling Salesman Problem using Ant Colony Optimization
byIRJET Journal
 
PDF
Travelling SalesMan Problem(TSP)
byAkshay Kamble
 
PDF
A COMPARISON BETWEEN SWARM INTELLIGENCE ALGORITHMS FOR ROUTING PROBLEMS
byecij
 
PPTX
Ant colony optimization in multiple travelling salesman problem
byParisha Arya
 
PDF
A heuristic approach for optimizing travel planning using genetics algorithm
byeSAT Journals
 
PDF
A heuristic approach for optimizing travel planning using genetics algorithm
byeSAT Publishing House
 
PDF
A New Approach To Solving The Multiple Traveling Salesperson Problem Using Ge...
byApril Smith
 
PPTX
PPT_Aritifical_Intelligence_Chapter4.pptx
byDSdivya12
 
PDF
Solving k-city multiple travelling salesman using genetic algorithm
byIAESIJAI
 
PPTX
Solving the traveling salesman problem by genetic algorithm
byAlex Bidanets
 
PDF
Generating Travel Itinerary Using Ant Collony Optimization
byTELKOMNIKA JOURNAL
 
PPTX
ABC-GSX:Hybrid method to solve TSP
bygauravon
 
PDF
Swarm Intelligence Technique ACO and Traveling Salesman Problem
byIRJET Journal
 
Travelling salesperson problem
byAkshayIngole17
 
Travelling Salesman Problem
byShikha Gupta
 
Operation research model for solving TSP
byDrGovindshaysharma
 
Traveling salesman problem__theory_and_applications
bySachin Kheveria
 
Traveling sales man problem research.pdf
bymoulayalih0101
 
Travelling salesman problem using genetic algorithms
byShivank Shah
 
Solving TSP using Hybrid GACO
byShruti Gandhi
 
A Survey of Solving Travelling Salesman Problem using Ant Colony Optimization
byIRJET Journal
 
Travelling SalesMan Problem(TSP)
byAkshay Kamble
 
A COMPARISON BETWEEN SWARM INTELLIGENCE ALGORITHMS FOR ROUTING PROBLEMS
byecij
 
Ant colony optimization in multiple travelling salesman problem
byParisha Arya
 
A heuristic approach for optimizing travel planning using genetics algorithm
byeSAT Journals
 
A heuristic approach for optimizing travel planning using genetics algorithm
byeSAT Publishing House
 
A New Approach To Solving The Multiple Traveling Salesperson Problem Using Ge...
byApril Smith
 
PPT_Aritifical_Intelligence_Chapter4.pptx
byDSdivya12
 
Solving k-city multiple travelling salesman using genetic algorithm
byIAESIJAI
 
Solving the traveling salesman problem by genetic algorithm
byAlex Bidanets
 
Generating Travel Itinerary Using Ant Collony Optimization
byTELKOMNIKA JOURNAL
 
ABC-GSX:Hybrid method to solve TSP
bygauravon
 
Swarm Intelligence Technique ACO and Traveling Salesman Problem
byIRJET Journal
 

Recently uploaded

PDF
The Journey Is The Reward - Apple Maps Travel Companion
byJohn Andrews
 
PDF
KV Caching Strategies for Latency-Critical LLM Applications by John Thomson
byScyllaDB
 
PDF
Rivian's Push Notification Sub Stream with Mega Filter by Marcus Kim & Saahil...
byScyllaDB
 
PPTX
Top Trends in Kubernetes Security: TalosCon 2025
byCheryl Hung
 
PDF
Getting the Best of TrueDEM – October News & Updates
bypanagenda
 
PDF
The invisible key: Securing the new attack vector of OAuth tokens
byGianluca Varisco
 
PDF
From Data Chaos to Copilot Confidence: A Step-by-Step Microsoft 365 Copilot R...
byNikki Chapple
 
PDF
What is Google Cloud Platform (GCP)? A 5-Minute Guide for Beginners (2025)
byTeleglobal International
 
PPTX
Session 5 - Specialized AI Associate Series: Build a Document Understanding ...
byDianaGray10
 
PPTX
"Towards React Server Components in Clojure", Roman Liutikov
byFwdays
 
PDF
Session 4 - Specialized AI Associate Series: UiPath Document Understanding O...
byDianaGray10
 
PDF
Unlock This Brilliant App Freezur That Builds Brainrot Videos That Captivate ...
bySOFTTECHHUB
 
PPTX
WUG-DMV: Workfront Wizards: Tips & Tricks Exchange (Sept 2025)
byWUG-DC MARYLAND VIRGINIA
 
PDF
Session 2 - Agentic Orchestration with UiPath Maestro
byDianaGray10
 
PDF
Expert Python App Development Company for Modern Enterprises
bySynapseIndia
 
PDF
UnityNet Digital Sovereignty Checklist 2025-10-13
byLHelferty
 
PDF
UiPath DevConnect 2025: Introduction to Agentic Automation
byDianaGray10
 
PDF
From Bots to Brains: Getting Started with Agentic Automation in UiPath
byUiPathCommunity
 
PDF
Draginoć•†ć“ă‚«ă‚żăƒ­ă‚° 2025.7 LoRaWAN・NB-IoT・LTE-M(LTE Cat.M1)ćŻŸćżœă‚»ăƒłă‚”ăƒŒăƒȘă‚čト
byCRI Japan, Inc.
 
PDF
Data Structure - 12 Graph
byAndiNurkholis1
 
The Journey Is The Reward - Apple Maps Travel Companion
byJohn Andrews
 
KV Caching Strategies for Latency-Critical LLM Applications by John Thomson
byScyllaDB
 
Rivian's Push Notification Sub Stream with Mega Filter by Marcus Kim & Saahil...
byScyllaDB
 
Top Trends in Kubernetes Security: TalosCon 2025
byCheryl Hung
 
Getting the Best of TrueDEM – October News & Updates
bypanagenda
 
The invisible key: Securing the new attack vector of OAuth tokens
byGianluca Varisco
 
From Data Chaos to Copilot Confidence: A Step-by-Step Microsoft 365 Copilot R...
byNikki Chapple
 
What is Google Cloud Platform (GCP)? A 5-Minute Guide for Beginners (2025)
byTeleglobal International
 
Session 5 - Specialized AI Associate Series: Build a Document Understanding ...
byDianaGray10
 
"Towards React Server Components in Clojure", Roman Liutikov
byFwdays
 
Session 4 - Specialized AI Associate Series: UiPath Document Understanding O...
byDianaGray10
 
Unlock This Brilliant App Freezur That Builds Brainrot Videos That Captivate ...
bySOFTTECHHUB
 
WUG-DMV: Workfront Wizards: Tips & Tricks Exchange (Sept 2025)
byWUG-DC MARYLAND VIRGINIA
 
Session 2 - Agentic Orchestration with UiPath Maestro
byDianaGray10
 
Expert Python App Development Company for Modern Enterprises
bySynapseIndia
 
UnityNet Digital Sovereignty Checklist 2025-10-13
byLHelferty
 
UiPath DevConnect 2025: Introduction to Agentic Automation
byDianaGray10
 
From Bots to Brains: Getting Started with Agentic Automation in UiPath
byUiPathCommunity
 
Draginoć•†ć“ă‚«ă‚żăƒ­ă‚° 2025.7 LoRaWAN・NB-IoT・LTE-M(LTE Cat.M1)ćŻŸćżœă‚»ăƒłă‚”ăƒŒăƒȘă‚čト
byCRI Japan, Inc.
 
Data Structure - 12 Graph
byAndiNurkholis1
 

Machine Learning Lecture 5 (Genetic Algorithm)

  • 1.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Learning Machine Learning :: Session 3 Partha Dey, PhD Dept. of Mechanical Engineering Academy of Technology 18th November 2018 1 / 10
  • 2.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Overview 1 Ant Colony Optimization 2 Genetic Algorithm Genetic Operators 3 The ACO/GA flowchart 4 The ACO/GA code 5 Acknowledgements 2 / 10
  • 3.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route 3 / 10
  • 4.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route 3 / 10
  • 5.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route 3 / 10
  • 6.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route 3 / 10
  • 7.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route She first 1 finds out the inter-city distances — 3 / 10
  • 8.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route She first 1 finds out the inter-city distances — Kol Ben Guw Che Mum Kol Ben 1560 Guw 524 2082 Che 1356 290 1884 Mum 1654 846 1034 2088 Source : www.distancecalculator.net 3 / 10
  • 9.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route She first 1 finds out the inter-city distances — Kol Ben Guw Che Mum Kol Ben 1560 Guw 524 2082 Che 1356 290 1884 Mum 1654 846 1034 2088 Source : www.distancecalculator.net . . . then 2 lists down all the 4 ! ways to plan her tour, 3 / 10
  • 10.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Salesman Problem (TSP) A salesperson has been given the assignment to visit . Bengaluru, . Guwahati, . Chennai and . Mumbai, from Kolkata and come back to her hometown. . . along the shortest route She first 1 finds out the inter-city distances — Kol Ben Guw Che Mum Kol Ben 1560 Guw 524 2082 Che 1356 290 1884 Mum 1654 846 1034 2088 Source : www.distancecalculator.net . . . then 2 lists down all the 4 ! ways to plan her tour, and finally 3 picks the shortest of them. 3 / 10
  • 11.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. 4 / 10
  • 12.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. 4 / 10
  • 13.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. BOM, 4 / 10
  • 14.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. BOM, BLR, 4 / 10
  • 15.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. BOM, BLR, MAA, 4 / 10
  • 16.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. BOM, BLR, MAA, GAU 4 / 10
  • 17.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance 4 / 10
  • 18.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. 4 / 10
  • 19.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . 4 / 10
  • 20.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . weaker untraversed trails . . . . . . 4 / 10
  • 21.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . weaker untraversed trails . . . . . . evaporate fast 4 / 10
  • 22.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . showing the way to an optimized path. . . 4 / 10
  • 23.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . showing the way to an optimized path. . . for our actual tranveller to follow 4 / 10
  • 24.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Travelling Ants to solve the TSP Now, the nature-loving salesperson has a few servants whom she asks to do the trial tours for her. As per order, 4 ants instantly start from Kolkata towards the four different cities. Each ant completes a tour and is Rewarded ∝ 1 tour distance They continue this quite a few number of times, each time depositing some pheromone proportional to its reward. The next ants follow stronger pheromone trail. . . . . . showing the way to an optimized path. . . for our actual tranveller to follow, without doing sort(4 ! sums) 4 / 10
  • 25.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) 5 / 10
  • 26.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3, 11) 5 / 10
  • 27.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (4, 6) 5 / 10
  • 28.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3.6, 7.5) 5 / 10
  • 29.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3.6, 7.5) Hence, we minimize A = 2πrh + 2πr2 subject to V = πr2 h ≄ 300 5 / 10
  • 30.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3.6, 7.5) Hence, we minimize A = 2πrh + 2πr2 subject to V = πr2 h ≄ 300 We ‘encode’ the two dimensions (r and h) in the form of a gene 5 / 10
  • 31.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem A can needs to be manufactured which is : cylindrical in shape should hold at least 300 ml liquid We will want to know the size of the can which will require the minimum material. Let the dimensions be denoted as (r, h) (3.6, 7.5) Hence, we minimize A = 2πrh + 2πr2 subject to V = πr2 h ≄ 300 We ‘encode’ the two dimensions (r and h) in the form of a gene 1 0 0 1 | 1 0 1 0 # Hackers : Any principle for encoding ? 5 / 10
  • 32.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem (solved with GA) Let’s take a possible range of values of r and h : r (coded) | h (coded) 2.5 7→ 0 0 0 0 | 5.0 7→ 0 0 0 0 2.6 7→ 0 0 0 1 | 5.5 7→ 0 0 0 1 ... 7→ ... ... | ... 7→ ... ... 4.0 7→ 1 1 1 1 | 12.5 7→ 1 1 1 1 So, any can dimension, say (3.6, 7.5) could be encoded as : 6 / 10
  • 33.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem (solved with GA) Let’s take a possible range of values of r and h : r (coded) | h (coded) 2.5 7→ 0 0 0 0 | 5.0 7→ 0 0 0 0 2.6 7→ 0 0 0 1 | 5.5 7→ 0 0 0 1 ... 7→ ... ... | ... 7→ ... ... 4.0 7→ 1 1 1 1 | 12.5 7→ 1 1 1 1 So, any can dimension, say (3.6, 7.5) could be encoded as : 1 0 1 1 | 0 1 0 1 6 / 10
  • 34.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Another Optimization Problem (solved with GA) Let’s take a possible range of values of r and h : r (coded) | h (coded) 2.5 7→ 0 0 0 0 | 5.0 7→ 0 0 0 0 2.6 7→ 0 0 0 1 | 5.5 7→ 0 0 0 1 ... 7→ ... ... | ... 7→ ... ... 4.0 7→ 1 1 1 1 | 12.5 7→ 1 1 1 1 or, say (4, 6) would be encoded as : 1 1 1 1 | 0 0 1 0 6 / 10
  • 35.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 1 1 1 1 0 0 1 0 7 / 10
  • 36.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 We choose a crossover point, 7 / 10
  • 37.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. 7 / 10
  • 38.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. 7 / 10
  • 39.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. → (3.6, 6) → (4, 7.5) 7 / 10
  • 40.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. 7 / 10
  • 41.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. 7 / 10
  • 42.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, 7 / 10
  • 43.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 1 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled 7 / 10
  • 44.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled 7 / 10
  • 45.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled (to avoid stagnation of solutions) 7 / 10
  • 46.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled (to avoid stagnation of solutions) These operators are called Crossover and Mutation. They mimic the natural evolution phenomena. 7 / 10
  • 47.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The Genetic Operators Let’s take these two encoded strings, viz. : 1 0 1 1 0 1 1 0 × 1 1 1 1 0 0 1 0 1 0 1 1 0 0 1 0 1 1 1 1 0 1 1 0 1 1 1 1 0 0 1 0 We choose a crossover point, across which we interchange genetic material among parents. Thus new solutions are born inheriting characteristics from both parents. Sometimes, a random bit is toggled (to avoid stagnation of solutions) In each iteration (or generation), better performing solutions are selected with higher probability. So offsprings improve. 7 / 10
  • 48.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The algorithm for ACO 8 / 10
  • 49.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The algorithm for GA 8 / 10
  • 50.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge The R code for GA CAN - ga(type = real-valued, fitness = function(can.dim) GA.obj(can.dim), lower = c(2.5, 5), upper = c(4, 12.5), popSize = 20, maxiter = 10) GA.obj - function(rh) { p - 3.1416; r - rh[1]; h - rh[2] V - p*r*r*h; A - 2*p*r*h + 2*(p*r*r) if (r ... | r ...) return(-100) if (h ... | h ...) return(-200) if (V 300) return(-300) return(1000 - A) } 9 / 10
  • 51.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements 10 / 10
  • 52.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee 10 / 10
  • 53.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya 10 / 10
  • 54.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil 10 / 10
  • 55.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity 10 / 10
  • 56.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana 10 / 10
  • 57.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana Prof. Ashoke Banerjee 10 / 10
  • 58.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana Prof. Ashoke Banerjee Prof. Shubhabrata Dutta 10 / 10
  • 59.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana Prof. Ashoke Banerjee Prof. Shubhabrata Dutta Prof. Kalyanmoy Deb and Prof. Marco Dorigo 10 / 10
  • 60.
    Learning ML Partha Dey, PhD AntColony Optimization Genetic Algorithm GA operators Flowcharts R Codes Acknowledge Without whom this course would be impossible Acknowledgements Prof. J. Banerjee Prof. D. Bhattacharya Prof. A. Sil Prof. D. K. Maity Prof. A. K. Rana Prof. Ashoke Banerjee Prof. Shubhabrata Dutta Prof. Kalyanmoy Deb and Prof. Marco Dorigo All of you with your tried-and-tested patience 10 / 10