Here is the link to download the particle simulation (without me of course :D ): au.mathworks.com/matlabcentral/fileexchange/69027-simulation-of-particles-in-particle-swarm-optimization
Never expected to find such a brilliant explanation in the first search attempt, doing literature review for Truss Optimization using PSO and found it to be extremely helpful. Thanks a lot!
Man you're the best!! I really didn't think I'd be able to finish my optimization homework and didn't think I could learn PSO so quickly but your explanation was so amazing I was really able to learn it by the end of the video and optimize my homework! Thank you SO MUCH!! You're amazing!
I've been using PSO since about 2005. Thank you for teaching it and keeping it alive. I still have my copy of "Swarm Intelligence" by Kennedy and Eberhart.
Hello. Please can I get your contact? I'm conducting a research on it. This is the topic: PSO-Based Optimization of Power System Stabilizers for Enhancing Small-Signal Stability
This was extremely helpful! I'm doing a research project on metaheuristics and I'm trying to familiarize myself with some of the various techniques that can be used. I'm very new to these heuristic concepts but this video made it very digestible for me. Thank you!
The improvement of a number of economic problems requires a good strategy, that is say to find a good algorithm. The choice of algorithm is a key point for success. Because no algorithm can solve all the problems of the same efficiency. We can define an algorithm as a succession of steps that lead to a solution for a given problem. Required work : -. Give the steps necessary to run the algorithm in question. - Using this algorithm on an example of your choice, give the steps followed and its programming
6:00 I believe that it is incorrect to state that he can sleep anywhere in the green shaded area, since the vector are all scaled scaled equally, if each of the three directions A, B, C are walked an equal distance d, then the total distance D is D = Ad + Bd + Cd = d (A + B + C), which means that no matter the value of d (0 km, 5km, 10km, 20km), the direction will be the same, namely (A+B+C)/||(A+B+C)||, which is the normalised value of (A+B+C). Therefore, all possible end points lie in the same direction. This is only changed once at 12:00 the variables w, c1, c2 are added to tune the contribution of each term.
Thanks for sharing. It really provides a clear explanation on a a really interesting concept. Great thanks for this. Only thing that bugs me, based on your visual illustration (around [7:30]) we see that the search space tend to overlaps both between and within search agents, 1. would you agree that it seems inefficient ? 2. How would the Pyhton code looks like to avoid mapping again the region already mapped in the previous search? 3. Finally why not simply use a parallel grid search which would avoid any overlap both within and across parallelised process ? Thanks again for sharing and looking forward to your reply Dr Mirjalili,Olivier
@ Ali Mirjalili I really like this alg. But I have a question, has there been any attempt to preserve the energy of the system? By that I mean, have particles near the local optima move more slowly and have those far away gain velocity "lost" by the slower moving particles. This need not be an exact preservation of energy, but I think it helps explain the idea: Those things that are far away move around faster exploring more broadly the unknown space while the portion that is closer focus on zeroing in on the exact best values
How can we add a velocity component [that has both displacement(distance) and direction] to a distance component? As we cannot add two components of different dimensions. Could you explain?
Why are we using random function in the equation?? Is it for the randomness of the particles used for observation or whether any specific reason is there for use of random function in the codes??
It was a brilliant video. Thanks Just one issue is not clear to me. How do you sum the distance with velocity as they are not from the same measures? Velocity needs to be multiplied by the time that has passed and then summed with the distances updated with regard to the personal best solution and the global best solution.
Thank you for the video, It help me quite a lot in my understanding of the PSO algorithm. PS : I would just like to point you that unless i missed some part, when you show the possible landing area, there seems to be a mistake unless r1 >= r2 >= r3 :/
Thank you very much, sir, this is by far the best video on PSO, sir please Is there some other meta-heuristic methods that have a lower computational time than the PSO?
Here is the link to download the particle simulation (without me of course :D ): au.mathworks.com/matlabcentral/fileexchange/69027-simulation-of-particles-in-particle-swarm-optimization
pretty vivid example, thx
may I use your simulation of POS part in my video
@@lingfengliu955 000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000⁰⁰⁰⁰00000000000⁰000⁰0⁰⁰000⁰00⁰⁰00000⁰000000000000000000000000000000000000000000000000000000000000000000000000000000000000000⁰00000000000⁰00000000000000000000000000000000000000⁰0000⁰000000000000000000000000⁰000000000000000⁰⁰0000⁰00000000⁰000000000000⁰00⁰0⁰000000⁰00000⁰00⁰0⁰000⁰0⁰0⁰00⁰⁰⁰0⁰00000⁰⁰⁰⁰⁰0ĺ00000000000000⁰⁰⁰00⁰000000000⁰000000⁰0⁰⁰⁰0000⁰⁰
Thank you sir, elaborately explained, you're Genius sir, thanks again.
Thanks for sharing such outstanding data
May your career converge in a global maximum, Sir.
Never expected to find such a brilliant explanation in the first search attempt, doing literature review for Truss Optimization using PSO and found it to be extremely helpful. Thanks a lot!
What Great content in 6 years ago! Thank you.
I think this is the best video on PSO on YT right now. Perfect mix of exploration of topics and exploitation/depth on each one.
this is by far the best video on pso i have came across.
I'm not in the field of PSO but the teaching in this video is crystal clear and illustrative. Great explanation even to a layman like me!
shoutout bro i love ur papers keep grinding 👆
Man you're the best!! I really didn't think I'd be able to finish my optimization homework and didn't think I could learn PSO so quickly but your explanation was so amazing I was really able to learn it by the end of the video and optimize my homework! Thank you SO MUCH!! You're amazing!
This video is really out of the world. Too good.
WOW !! it was by far the best video for PSO. tnx
This 20min video was a 2 hour lecture at my university, still this explains it better.
The instructor does a good job introducing the concept of PSO and explaining the role of each component involved in the base algorithm
I've been using PSO since about 2005. Thank you for teaching it and keeping it alive. I still have my copy of "Swarm Intelligence" by Kennedy and Eberhart.
do you have the code of PSO?
Hello.
Please can I get your contact?
I'm conducting a research on it.
This is the topic: PSO-Based Optimization of Power System Stabilizers for Enhancing Small-Signal Stability
Professional and easy explanation with examples...we need YT channels like this
The best explaination of PSO i've found yet!
Wow! Didn't expect the video to be so informative and visually appealing. This is the best explanation I found on this topic so far, thanks a ton!
Best explanation of PSO I have come across so far!
this might be the best explanation ever made about this subject. Thank you so much
The video title was not a lie. Great video!
your explanation really easy to understand.. thank you
My goodness! This is by far the best presentation I have come across! Thank you so much!
Amazing video and the best explanation of PSO found on TH-cam.
Thanks Ali Mirjalili
It's my first look into PSO, glad that I found this. Thanks a lot!!!
The best tutorial on PSO on youtube..
This was extremely helpful! I'm doing a research project on metaheuristics and I'm trying to familiarize myself with some of the various techniques that can be used. I'm very new to these heuristic concepts but this video made it very digestible for me.
Thank you!
Check out the Demon Algorithm
The improvement of a number of economic problems requires a good strategy, that is
say to find a good algorithm.
The choice of algorithm is a key point for success. Because no algorithm can solve
all the problems of the same efficiency.
We can define an algorithm as a succession of steps that lead to a solution for a
given problem.
Required work :
-. Give the steps necessary to run the algorithm in question.
- Using this algorithm on an example of your choice, give the steps followed
and its programming
A great video with a clear explanation. A great primer to the topic. Thank you !
I'm looking at tuning parameters of ANN model by using PSO technique, and so far this was the best explanation of PSO, thanks.
hi , what kind of results did u got?
Very well explained, Sir Ali! Thank you for this video!
I usually use GA to solve optimization problem but I will learn PSO. Thanks for the video.
do you mean Genetic algorithm ? and if yes how do you do it?
@@Manish-rq9lr Yes, I do
yeah both algorithms are good
@@LEARNWITHPANDA thanks buddy
Really nice and clear explanation! The tests were fundamental to understand how the algorithm works! Thanks a lot!
🐐 - very clearly explained and the visuals helped a lot to reinforce the concepts.
6:00 I believe that it is incorrect to state that he can sleep anywhere in the green shaded area, since the vector are all scaled scaled equally, if each of the three directions A, B, C are walked an equal distance d, then the total distance D is D = Ad + Bd + Cd = d (A + B + C), which means that no matter the value of d (0 km, 5km, 10km, 20km), the direction will be the same, namely (A+B+C)/||(A+B+C)||, which is the normalised value of (A+B+C). Therefore, all possible end points lie in the same direction.
This is only changed once at 12:00 the variables w, c1, c2 are added to tune the contribution of each term.
*Thank you Sir*..*Concise and Lucid*...really beneficial
Thanks Darshan
This was an amazing video. Extremely clear and informative. Thank you so much.
I'd really like to thank you for making it understandable
BEST PSO VIDEO SO FAR
Very well done visualization, especially the effects of the parameters :) Thanks!
Thanks :)
This is really an informative video I was looking for. Thank you so much Ali
brother i saw your name in your research article today i saw you thank you so much
Thank you sir. This is probably the best video to understand PSO.
Even I understand PSO now, so this must be a great tutorial!
Amazing bro, very cool explanation,I've loved when particles started looking for your hand
A really great explanation ,with a good diagrammatic examples.
Very good pronunciation and illustrative examples supported by mathematical arguments. Good Job!
great video! thank you tony stark!
Thanks! This is the best explanation on PSO!
Thank you 😊
I want to become well versed and educated like you sir. That's my wish
This is absolutely brilliant, thanks Teach!
The video is nice for the beginner. Thank you so much!
Best explanation ever !! Thanks.
Thanks for sharing. It really provides a clear explanation on a a really interesting concept. Great thanks for this. Only thing that bugs me, based on your visual illustration (around [7:30]) we see that the search space tend to overlaps both between and within search agents,
1. would you agree that it seems inefficient ?
2. How would the Pyhton code looks like to avoid mapping again the region already mapped in the previous search?
3. Finally why not simply use a parallel grid search which would avoid any overlap both within and across parallelised process ?
Thanks again for sharing and looking forward to your reply Dr Mirjalili,Olivier
Simple and comprehensive.Thanks
@
Ali Mirjalili
I really like this alg. But I have a question, has there been any attempt to preserve the energy of the system?
By that I mean, have particles near the local optima move more slowly and have those far away gain velocity "lost" by the slower moving particles. This need not be an exact preservation of energy, but I think it helps explain the idea: Those things that are far away move around faster exploring more broadly the unknown space while the portion that is closer focus on zeroing in on the exact best values
What an excellent explanation!
Crystal clear. Thanks to your excellent quality video!
What an explanation! Thank you so much.
How can we add a velocity component [that has both displacement(distance) and direction] to a distance component? As we cannot add two components of different dimensions. Could you explain?
Awesome sir...finally understood..thanks a lot sir
@Ali Mirjalili - Good Job , tried to explain a difficult topic in an easy way , Thanks.
Thanks Manish.
do you have explanation for sine cosine as well? Ive read your paper but still cannot understand
Incredible explenation! Definitely going to look into your Udemy courses seeing the quality of this video. Thank you!
مهندس جان خداقوت خیلی مفید بود
Thank you so much for this explanation. It is appreciated.
What a nice explanation! Heads up for your effort
Thank you!. crystal clear!
Very nicely explained, great video :))
Dear Ali Mirjalili,
Thank you for this valuable presentation, we would like to know if you can do the dame for GWO.
Yassine Chaibi Hi. Thanks for your kind message. I have a course on GWO.
Very good video - the best on PSO
it was very nicely described Ali. It is appreciated
Very informative and good illustration...
That was bloody fantastic mate
Thank you. This video is very easy to understand!
great job, your explanation is fantastic
Amazing !Crystal clear
Amazing description. Thanks a lot.
Thank you for the video, It help me quite a lot in my understanding of the PSO algorithm.
can someone please explain to me what is the difference between PSO and simulated annealing? thanks
Brilliant Explanation..
Why are we using random function in the equation??
Is it for the randomness of the particles used for observation or whether any specific reason is there for use of random function in the codes??
This is a great video. Thank you for sharing
Great explanation... Good job Sir.
It was a brilliant video. Thanks
Just one issue is not clear to me. How do you sum the distance with velocity as they are not from the same measures? Velocity needs to be multiplied by the time that has passed and then summed with the distances updated with regard to the personal best solution and the global best solution.
Excellent video! It is really helpful!
Dear Ali, why is it beneficial to have the walking distance randomised? I don't see the problem of doing the same with fixed 10km walks
Thank you for the video, It help me quite a lot in my understanding of the PSO algorithm.
PS : I would just like to point you that unless i missed some part, when you show the possible landing area, there seems to be a mistake unless r1 >= r2 >= r3 :/
Great video, awesome visualizations :D
Great video man, thanks
Well explained. Thanks.
Thank you Dr Ali, You are a great teacher
What I don't understand is: how does the team now they are nearer to the treasure? Using try an error?
Very precise and informative
This was really cool bro, I loved it
really good, informative video
Thank you very much, sir, this is by far the best video on PSO, sir please Is there some other meta-heuristic methods that have a lower computational time than the PSO?
thanks man , it was really helpfull to watch your video .
Hi Ali. I liked the video. Thank you for sharing!
Thank you sir for this good explain
Wow that is Very good explanation sir. Superb