Reddit reviews Convex Optimization
We found 4 Reddit comments about Convex Optimization. Here are the top ones, ranked by their Reddit score.
We found 4 Reddit comments about Convex Optimization. Here are the top ones, ranked by their Reddit score.
I'll just leave these two quality references here for those of you that care about these branches of mathematics:
https://www.amazon.com/Convex-Optimization-Stephen-Boyd/dp/0521833787/ref=pd_lpo_sbs_14_img_1?_encoding=UTF8&psc=1&refRID=TVGGPQ59DZ58S2XSXMGD
https://www.amazon.com/Combinatorial-Optimization-Algorithms-Complexity-Computer/dp/0486402584
And for those of you who like to throw in a little probability theory to the mix for more real-world situations...
https://www.amazon.com/Introduction-Stochastic-Programming-Operations-Engineering/dp/1461402360
If you are interested enough in machine learning that you are going to work through ESL, you may benefit from reading up on some math first. For example:
Without developing some mathematical maturity, some of ESL may be lost on you. Good luck!
Hi OP,
I found myself in a similar situation to you. To add a bit of context, I wanted to learn optimization for the sake of application to DSP/machine learning and related domains in ECE. However, I also wanted sufficient intuition and awareness to understand and appreciate optimization it for it's own sake. Further, I wanted to know how to numerically implement methods in real-time (embedded platforms) to solve the formulated problems (Since my job involves firmware development). I am assuming from your question that you are interested in some practical implementation/simulations too.
​
< A SAMPLE PIPELINE >
Optimization problem formulation -> Enumerating solution methods to formulated problem -> Algorithm development (on MATLAB for instance) -> Numerical analysis and fixed-point modelling -> Software implementation -> Optimized software implementation.
&#x200B;
So, building from my coursework during my Masters (Involving the standard LinAlg, S&P, Optimization, Statistical Signal Processing, Pattern Recognition, <some> Real Analysis and Numerical methods), I mapped out a curriculum for myself to achieve the goals I explained in paragraph 1. The Optimization/Numerical sections of the same is as below:
&#x200B;
OPTIMIZATION MODELS:
NUMERICAL METHODS:
&#x200B;
Personally I think this might be a good starting point, and as other posters have mentioned, you will need to tailor it to your use-case. Remember that learning is always iterative and you can re-discover/go deeper once you've finished a first pass. Front-loading all the knowledge at once usually is impractical.
&#x200B;
All the best and hope this helped!
Some other references for aerospace applications:
And for convex optimization in general:
Also for freely available software there is IPOPT, PSOPT and ICLOCS that I'm aware of:
For commercial software there is SNOPT and GPOPS-II: