Good morning everyone,
I work on the optimization of solar thermal power plant using ANN with Matlab software, to obtain the minimum value of the levelized cost of electricity (LCOE), I already simulate some cases (100 case) for using the inputs and output (target) in this ANN.
My question is which are the next steps to obtain the min value of LCOE ?
Thanks in advance.
Best regards.
ANN coupled genetic algorithm can be used for obtaining the same. GA tool box available in matlab tool box can be made use....
Thank you very much for your prompt reply, and for your time.
Hi,
as Rajesh wrote You can use the genetic algorithm toolbox to find the minimum. There is good documentation in Matlab website to do this including help with local and global Optimums. There is also possible to use multiobjective optimization if you for example would like to optimize CO2emissions.
Dear Taqiy,
There are so many optimization methods and tools to solve your problem. GA and PSO are popular among university students but they are meta-heuristic and thus you cannot be sure the result is the absolute optimum.
As an engineer, I recommend you using GAMS software which has so many solvers able to optimize your problem based on mathematical rules and not meta-heuristically.
See http://www.gams.com/
Regards,
Morteza Shabanzadeh
http://mortezash.blogfa.com/
Most people here answered correctly your question, but they did not touch what I believe to be of utmost importance: never try to optimize black box models. A neural network will only be as good the training data, and since any optimization (PSO, GA, etc) tool you use can go anywhere in the search space, you'll have a very significant risk going where there was no training data and have a meaningless result. The neural network can only be used with input data that came from the same "source" as the training data, it will not extrapolate well. I myself have tried to do this and failed, so heed my warning.
- Badges
- Science topic