You are a legend , you are a diamond , you are a hero even I stay here for the whole of my life I can't describe how a wonderfull teacher realy you are , thank you so much .
I believe this video is very helpful, expecially for those people approaching ACO for the first time. Having said that, I have a question about ACO. The heuristic information in your example is based on the cost of the single arc. But in many application the cost of a single component is influenced by the value assumed by the other componets. For istance, consider a problem of pump scheduling in water distribution networks, the actuation of a pump in a precise time interval produce a cost that depends on the pump state (on/off) during the preceding time intervals, and it influences the costs of the subsequent time intervals. In this case, how the pheromone trails can be updated? Because the cost function depends on the whole path the ant follows and not by a single component of the path.
Incredibly well explained. Especially with the visual and mathematical examples. Thank you very much Mister Mirjalili! I took the opportunity to subscribe to some of your udemy course. I hope, later this year, I will have enough time to learn more about this and the other algorithms.
I really enjoyed this clearest way to explain things and making super simple and understandable. Thank you so much. this my first sight of your course and i will stay learning here
This lecture remained very use full. Thank you so much sir. Kindly upload a video, regarding how to program this algorithm (ACO) in any of the programming language like python, IBM LOG cplexx or matlab. Thanks
The clearest video on youtube about Ant Colony Optimization, thank you very much Dr. Ali. Just a little correction on [5:05](ua-cam.com/video/783ZtAF4j5g/v-deo.html), the formula for the total pheromone from edge i-j, (T^{i, j}) shouldn't have the K superscript.
Thank you, Ali.. your video is very helpful to finish my task. I didn't know before, this algorithm can be understood in an easy way. Hope to see more of your video. Thank you
amazing video!!! very well explained!! Only suggestion, not trying to be rude, but some words at moments can be a tiny bit hard to understand, my suggestion is maybe type of the captions? Other than that great video!
Reminds me on Markov chains. Hard part is that every time concentration of feromon is different so probability changes. If also quality of road change I wonder how to calculate stationary distribution if there is any at all.
I have a question concerning evaporation rate - Why is time not a variable in the equation for pheromone levels? Additionally, how can one include more environmental variables such as temperature, terrain, humidity, light, etc. or colony variables such as speed of ants, density, distribution, type of ant, etc. into the algorithm? Is there any other type of communication that occurs between individuals?
Shouldn't the kth tau equal the old tau plus the change in tau, if there is no vaporization? The way it is written now (6:10) means rho = 1, which equals maximum vaporization. Or am I misunderstanding something?
guys please when we are talking about I,j, Is the current ant position and j is one of the next steps around it like the left to right ? or j is the final destination like nest or car and tree?
Wanna know whether the pheromone is updated locally on each arc visited by ant or the pheromone is updated globally after the ants visits all the arcs in a complete path....?
If my problem is a combinatorial problem, but I can only calculate the cost once a full trail has been formed, how can I calculate the distance between two nodes? Or should I just ignore that part?
you have upload your code inmatlab . i ma facing difficulty in running that code in matlab. can you help me in that . i tink i am facing a problem i calling function into other function in matlab. can you give some advise
shouldn't the evaporation be subtracted rather than added to current pheromone level? at 10:44 the pheromone level has increased from 0.1 to 0.6 with evaporation.
at 6:44 in the video, you say if roh is equal to 1 the evaporation is at its maximum, but mathematically if roh is equal to 1 the first term of second equation will become zero and we will be left with the equation which is equal to pheromone level without evaporation. so what's going on here?
Sir, thank you for the effective video, and for your effort, can you please explain to us the Artificial Bee Colony optimization, sir can I use the ABC algorithm to solve the shortest path problem? another question please, what are the meta-heuristics methods used to solve the shortest path problem? waiting for your response thank you in advance.
Hi Ali , First of all thank you for this amazing video and , my question is , how we can choose the destination city from a bunch of cities with equal probability values?
There are different models. Some, chose the closest city. Others, simply don't care. The idea is for the simulation to, eventually, show you the optimum path. So even when two have the same probability for a given time, in the long term, the one which favors the best path will end up with a higher value. If two paths are equally optimum, then your model - having same probabilities - will reveal it.
Ali, I've watched many UA-cam videos and yours is without a doubt the best explanation. I'm working on a small project using your algorithm to develope an app for iOS. At 9:21, what if there is a path that hasn't been taken yet? That path would have a Pheromone value of 0? Even with the evaporation model the value would be 0. How do you consider this path to be taken? Thank you.
I assume they would normally originate from a nest, but in reality there will of course be situations where scouts take random routes that eventually become solid paths. That's why most ant lines aren't the perfect shortest path from source to nest, but are shorter relative to routes taken by other scouts.
If you intend to use ACO for model optimization then you need to implement each model as an ant & use the idea of pheromones either as penalty or dependent on your desired parameters. By doing so, you essentially put the models as ensemble & are optimizing the ensemble by favoring the best model. Having said that, it's not essential to implement ACO for optimization of predictive model control courtesy of requirements of high processing cost & nearly chance of not giving the correct output (ACO is notorious for not always giving best results). Thus, i suggest looking at other optimization methods or look into cost/error functions in terms of penalty for ensemble methods.
You are definetely what students want, pure and steady explained knowledge. Keep it up we really appreciate your work here !!
crisp and clear, everything would be easy, when there are teachers like you
You are a legend , you are a diamond , you are a hero even I stay here for the whole of my life I can't describe how a wonderfull teacher realy you are , thank you so much .
Thanks, the diamond is in your kind heart :)
You explained the complex ACO problem in such simple understandable way! Great job!
the best explanation of ACO ever
great work
specially the slides are too good to understand , thank you brother
I believe this video is very helpful, expecially for those people approaching ACO for the first time. Having said that, I have a question about ACO. The heuristic information in your example is based on the cost of the single arc. But in many application the cost of a single component is influenced by the value assumed by the other componets. For istance, consider a problem of pump scheduling in water distribution networks, the actuation of a pump in a precise time interval produce a cost that depends on the pump state (on/off) during the preceding time intervals, and it influences the costs of the subsequent time intervals. In this case, how the pheromone trails can be updated? Because the cost function depends on the whole path the ant follows and not by a single component of the path.
At first I was having hard time understanding but then everything made so much sense ... that was some awesome explanation ... !!! good work man !
thank you for your crystal and clear description of ACO. I hope you become hero of your life
one of the best videos about ACO I have ever seen thanks a lot
The best teacher......period
Very, very good explanation and numerical examples helped me a lot to understand it even more. Amazing job!
Great and subtle explanation. The numerical examples and graphics are incredibly helpful. Thank you so much!
Thanks a million for this video. You saved my grades
The content is really helpful and even the formatting of the video is very sorted. Thanks
wowwwwww mannnnnnnn, just god level teaching .god bless u bro
Incredibly well explained. Especially with the visual and mathematical examples. Thank you very much Mister Mirjalili!
I took the opportunity to subscribe to some of your udemy course. I hope, later this year, I will have enough time to learn more about this and the other algorithms.
I really enjoyed this clearest way to explain things and making super simple and understandable. Thank you so much. this my first sight of your course and i will stay learning here
This lecture remained very use full. Thank you so much sir.
Kindly upload a video, regarding how to program this algorithm (ACO) in any of the programming language like python, IBM LOG cplexx or matlab. Thanks
Thanks a lot. As far as I know, ACO has been implemented in a large number of programming languages.
Perfect video explaining ant colony optimization. Thank you indeed.
The clearest video on youtube about Ant Colony Optimization, thank you very much Dr. Ali. Just a little correction on [5:05](ua-cam.com/video/783ZtAF4j5g/v-deo.html), the formula for the total pheromone from edge i-j, (T^{i, j}) shouldn't have the K superscript.
Wonderful explanation. All doubts have been cleared.
Excellent visualization and presentation. Kudos
Thank you so much Dr. Ali. Its quite clear tp proceed with an application. Huge respect for you. Really great!
You have great teaching and presentation skill. Thank you for making this video.
Sir, please explain Antlion Optimization. I already emailed you about it. Please explain it with its mathematical function.
Thank you, Ali.. your video is very helpful to finish my task. I didn't know before, this algorithm can be understood in an easy way. Hope to see more of your video. Thank you
amazing video!!! very well explained!!
Only suggestion, not trying to be rude, but some words at moments can be a tiny bit hard to understand, my suggestion is maybe type of the captions? Other than that great video!
Great explanation, thank you very much!
Reminds me on Markov chains. Hard part is that every time concentration of feromon is different so probability changes. If also quality of road change I wonder how to calculate stationary distribution if there is any at all.
Is thia guy the same Mirjalili who gave the fireflies,moth and crow optimization algorithm 😲😲...u are so awsm
Thank you for the effective video 🤩.
You made me clear so that I can handle my mini project!!!
Hi Nitheesha shree i am working on route optimization. can you please share your project.
many thanx , amazing presentation.
I have a question concerning evaporation rate - Why is time not a variable in the equation for pheromone levels?
Additionally, how can one include more environmental variables such as temperature, terrain, humidity, light, etc. or colony variables such as speed of ants, density, distribution, type of ant, etc. into the algorithm? Is there any other type of communication that occurs between individuals?
Thank you very much. It is clear. I understand it very well
You are a great teacher, thank you so much !!!!!
Shouldn't the kth tau equal the old tau plus the change in tau, if there is no vaporization? The way it is written now (6:10) means rho = 1, which equals maximum vaporization. Or am I misunderstanding something?
You are right, this should be written as : tau = ( 1 - rho ) .* tau;
Wow, that made it so much clearer, thank you!!
Thanks you so much sir for this wonderful explanation...this help me alot🙏🙏
Thank you very much Dear Dr.Ali for this video
guys please when we are talking about I,j, Is the current ant position and j is one of the next steps around it like the left to right ? or j is the final destination like nest or car and tree?
Please upload videos on particle swarm optimisation
Sir I want implementation of exam time table problem using ant colony
This helped me a lot. Thank you very much
Wanna know whether the pheromone is updated locally on each arc visited by ant or the pheromone is updated globally after the ants visits all the arcs in a complete path....?
very informative lecturer may be in future i will take some help regarding ACO
the pheromone matrix depends on the length of the solution or the number of parameters of the problem ?
Man, you are THE BEST !
If my problem is a combinatorial problem, but I can only calculate the cost once a full trail has been formed, how can I calculate the distance between two nodes? Or should I just ignore that part?
Great explanation
Thank you for the detailed explanation and the wonderful accuracy in explaining this algorithm. I hope you explain the gorilla algorithm as well.
What is the time complexity of any optimization based evolutionary algorithm?
you have upload your code inmatlab . i ma facing difficulty in running that code in matlab. can you help me in that . i tink i am facing a problem i calling function into other function in matlab. can you give some advise
Thanks for sharing the video.. understood the ACO
shouldn't the evaporation be subtracted rather than added to current pheromone level? at 10:44 the pheromone level has increased from 0.1 to 0.6 with evaporation.
Brilliant! I am wondering if you can use Bayes theorem for probability? Since you can observe the paths taken by the ants.
thanks , that was quite helpfull .
at 6:44 in the video, you say if roh is equal to 1 the evaporation is at its maximum, but mathematically if roh is equal to 1 the first term of second equation will become zero and we will be left with the equation which is equal to pheromone level without evaporation. so what's going on here?
Great video sir, can i use this model for optimization in scheduling of activities in a construction industry????
How we can apply ant colony with clustering?
Sir, thank you for the effective video, and for your effort, can you please explain to us the Artificial Bee Colony optimization, sir can I use the ABC algorithm to solve the shortest path problem? another question please, what are the meta-heuristics methods used to solve the shortest path problem? waiting for your response thank you in advance.
How do you guarantee that the pheromone on the edge will not become 0?
Superb..fida hogaya..want to register for the course..getting problem in sign up..can u provide lecture on firefly algorithm please
Sir,
Is there any lecture on Sine-cosine algorithm, that will explain the working principle clearly?
Nicely explained.....
The link in description is not working......
Thanks a lot. Ooops; I have updated the link.
I like your video, it is very clear. thanks a lot
please sir how to determine the matrix of pheromone
why do we need to connect every node with every other node???
Hi Ali , First of all thank you for this amazing video and , my question is , how we can choose the destination city from a bunch of cities with equal probability values?
There are different models. Some, chose the closest city. Others, simply don't care.
The idea is for the simulation to, eventually, show you the optimum path. So even when two have the same probability for a given time, in the long term, the one which favors the best path will end up with a higher value.
If two paths are equally optimum, then your model - having same probabilities - will reveal it.
@@anibaldk many thanks . I got it.
The last equation at 6:41 is wrong
Yep, it is supposed to be tau = ( 1 - rho ) .* tau;
shouldn't the vaporization be quadratic ? since its spreading in 2d space
great explanation!
Sir any multi objective optimization intergrated with artificial intelligence course on udemy?
amazing video
Ali, I've watched many UA-cam videos and yours is without a doubt the best explanation. I'm working on a small project using your algorithm to develope an app for iOS.
At 9:21, what if there is a path that hasn't been taken yet? That path would have a Pheromone value of 0? Even with the evaporation model the value would be 0. How do you consider this path to be taken?
Thank you.
I suppose that all the pheromones at the beginning will be set to 1
Thanks its very useful, please i want to know about a mathematical model of social spider optimization
I have not used social spider optimization
Can i use it to find area have more children but not have school??!
Hi Professor, I was trying to look for this original math model in an official published resource... What is your resource please can you share it?
Excellent ....Thankyou Bro
very good thank you so much
Sir any good multi objective course on udemy?
Thanks a lot.
brilliant man
How do we find and solve alpha and beta in that equation ? Please help me
It's up to you. You can choose any. Just try different values and leave values that work the most effective.
Great vid! Voted!
Hi Prof. Ali
Thanks for your explaining ACO.
How can I find your email?
How can i get these slides??
why do green ants choose that path? not others. there are 3 combinations using those 4 points. what is the reason?
where you get pheromone value without length of ant tour?
could you make a video that explains the theory and the code of bees algorithm and we will be able to find it in udemy,
Very clear
Thank you !!!
can you resolve it using python?
Do the ants have to be initialized from the same starting point? Or different?
Cause I'm seeing both online
I assume they would normally originate from a nest, but in reality there will of course be situations where scouts take random routes that eventually become solid paths. That's why most ant lines aren't the perfect shortest path from source to nest, but are shorter relative to routes taken by other scouts.
Thanks a lot! :D
Thanks a lot.
what does the alpha and beta defins in that particular equation? please help me .
Those are some parameters to change the impact of pheromone and the quality of an edge on calculating the probability.
Thanks alot Sir
how we can relize ant algaritham in ns2 simulater.can you help me ?
professor, great explanation, can i use this algorithm in model predictive control ? if yes, can you tell me how ?
If you intend to use ACO for model optimization then you need to implement each model as an ant & use the idea of pheromones either as penalty or dependent on your desired parameters. By doing so, you essentially put the models as ensemble & are optimizing the ensemble by favoring the best model. Having said that, it's not essential to implement ACO for optimization of predictive model control courtesy of requirements of high processing cost & nearly chance of not giving the correct output (ACO is notorious for not always giving best results). Thus, i suggest looking at other optimization methods or look into cost/error functions in terms of penalty for ensemble methods.
I want to make project on Grasshopper Optimization algorithm please give me some ideas...
Great video!