Numerical Optimization Algorithms: Constant and Diminishing Step Size

[@ChristopherLum]

In case it is helpful, all my optimization videos in a single playlist are located at kzbin.info/aero/PLxdnSsBqCrrHo2EYb_sMctU959D-iPybT. You can support this channel via Patreon at www.patreon.com/christopherwlum. Please let me know what you think in the comments. Thanks for watching!

[@sinarezaei218]

you are one of the best professors I have ever seen thanks for your videos

[@based3765]

Your channel is truly a gem. I was looking for a tutorial on Matlab Simulink with Arduino, ended up watching your math, biking, hiking and fridge videos, and I absolutely love it! Thank you very much, keep up with this manly content!

[@ChristopherLum]

Thanks for the kind words, I'm glad you enjoyed the video. If the find the these videos to be helpful, I hope you'll consider supporting the channel via Patreon at www.patreon.com/christopherwlum. Given your interest in this topic, I'd love to have you a as a Patron as I'm able to talk/interact personally with all Patrons. Thanks for watching!

[@GiovanniRiggi-j2q]

AE 501 - Johnny Riggi: It is interesting to visualize how the optimizations can change based upon the input parameters. Great lecture!

[@elijahleonen198]

AE501: The MATLAB graphs with the step sizes really helped me understand the content. Thank you for the great lecture!

[@Chuan-YuTsai]

AE501: Thanks for the example and for explaining two simple step size selection methods for numerical optimization algorithms.

[@maggiesplantgirl]

AE501: Great visuals as always! -Maggie Shelton

[@tonykuenzli2182]

AE501 - The demos and examples were really good at highlighting the advantages and pitfalls of these two techniques for picking a step-size

[@nathanlipshutz3140]

AE501. This is a topic that I've always been interested in exploring but never had the opportunity. I'm glad we're taking a look at it in this class. I'm excited to play with these numerical methods.

[@jacobgivens2000]

AE501: I enjoyed the visuals through the graphs generated explaining the step size. Great lecture!

[@ChristopherLum]

I am glad you found the graphs helpful in illustrating the difference between constant and diminishing step size.

[@calebwatson6331]

AE501: The visualization was super helpful in this, great demos!

[@ahmedashmaig]

AE501: The matlab examples are always super helpful!

[@enigma5627]

AE 501: Jesse Perez - this numerical approximation for calculating the step size, either constant or with a diminishing approach, reminds me of the robot vacuum cleaner. The robot takes enough steps to map your floor plan and eventually clean but we can confuse it if we put obstacles in its way.

[@LilanieAlfredaAbdur-Rahman]

Ae 501: Nice demonstration! Are the calculations, if done by hand (just for 2 or so iterations) as straight forward as they appear?

[@KennethWright-k2h]

AE501: Thank you for comparing methods to see the pitfalls of each.

[@noviv]

AE501: The more I watch the more this problem feels like weighted pathfinding in computer science - I wonder if techniques could be pulled from A* or other well-established pathfinding algorithms to make these numerical optimization algorithms more efficient.

[@jasonfranklin4614]

AE501: It interesting that optimization, as I watch this series, seems to really offer a very diverse number of challenges (even though I'm only scratching the surface here). So many options for solving problems, it would seem like a custom solution generally is the most optimal.

[@thefoxcatch]

Could you share the definition of 'diminishing' in your matlab code? Or is it as defined in the whiteboard, i.e., || grad( f(xK) ) || * gamma? I'm assuming gamma is a scalar constant...

[@thefoxcatch]

duh , gamma is fixed

[@cupdhyaya]

Please discuss particle swarm for constrained optimization.

[@ChristopherLum]

Hi Chandra, Thanks for reaching out. If you have questions or would like to request a video, I hope you'll consider supporting the channel via Patreon at www.patreon.com/christopherwlum. I'd love to have you as a Patron as I'm able to talk/interact personally with Patrons. Thanks for watching! -Chris

[@ThienNguyen-vx1td]

AE 501: Interested topic, and would like to do some lab with this. for that hands on experience.

[@AurashFilsoof]

AE: 501 Thank you

[@malachimorris1649]

Ae501 - Malachi Morris

[@ardacetken9090]

AE501: Arda Cetken

[@CodySmith]

AE501 Cody Smith

[@changlongwang5201]

[A E 501 student] watched - CW

[@TriMartz-f4x]

[AE501] choosing a step size seems a bit of trial and error and maybe some luck