Grasshopper

algorithmic modeling for Rhino

Built Social Network - The Digitalization of Participatory Processes
Bachelor-Thesis written in the wintersemester 2013/14 at the msa | münster school of architecture in the department urban planning.
Supervisor: Prof. Joachim Schultz-Granberg

Views: 400

Comment

You need to be a member of Grasshopper to add comments!

Comment by Igor Mitrić Lavovski on February 10, 2015 at 3:47pm

Niiiiceeeee! 

Comment by Martin D. on February 8, 2015 at 4:36am
Comment by Martin D. on January 11, 2015 at 10:12am
Klar! Ich werd mich Ende des Monats drum kümmern, dann hab ich ein wenig Zeit.
Hopefully I will also have the time, to work on the english short version.
Comment by RWNB on January 11, 2015 at 6:20am

hallo martin, interessante arbeit.

ich kann auch deutsch lesen, gibt es eine möglichkeit deine arbeit zu lesen?

grüsse

Comment by Mostapha Sadeghipour Roudsari on January 4, 2015 at 2:27pm

Thank Martin for the explanation. In particular what you mentioned in the last paragraph is pretty interesting. Best of luck and looking forward to see more of your work.

Comment by djordje on January 4, 2015 at 4:03am

Very interesting project! Would love to see the English version of it too.

Comment by Martin D. on January 4, 2015 at 3:43am
I would like to share it, but it is written in german. When I find some time, I will prepare some pics or a video about the technical background and let you know then!
What I can say atm is, that I used Sunray-Vectors for the Light-Objective in a way like Ladybug does it and a similar way for the Sightlines and Noise. Just drawing a line from every "window" to the yard, city, greenspace or for noise towards the street and checking if the line hits an obstacle.
For the best option I tried out different evolutionary solvers and ended up using octopus, since it gave me the best results. But I didn't use multi-objectives, I just calculated the fitness in a way like this: x1*Light+x2*Noise+(x3*View-yard+x4*View-green+x5*View-city)/3
x1-5 are all numbers between 0 and 1 so with them you set the importance of the factors and Light, Noise, View are interopolated, so that their maximum value is 1 and the lowest 0.

As you can see in the video, the performance of the different generated buildings does not differ so much. It would be more interesting to combine the apartment and building generation. But that was not possible, because of huge calculation times and an immense higher solution space.

Generally I would not do it again in this way, since it is really hard to give everyone their wanted apartment with the right size and location. People may have too high expectations which the outcome can't serve. More interesting for participatory processes could be Gaming, since people interfere more and can negotiate better. The programm is then giving feedback about sunshading, visibility etc.
That is what I am intersted in now.

Hope that helped already, feel free to ask more or just write me a pm ;-)
Comment by Mostapha Sadeghipour Roudsari on January 4, 2015 at 1:51am

Very nice Martin! Is there any chance that I can read your thesis? I'm particularly interested to see how did you set up the objective functions, and how did you end up selecting the best option(s) for an n-dimensional problem.

Comment by Bene on January 3, 2015 at 4:36pm

excellent work !

Comment by Martin D. on January 3, 2015 at 9:41am
I would like to give a big thanks to the Grasshopper-Community. I didn't really use so many Add-Ons, although I tried out many, like Octopus, Goat, ExcelReader, Horster, Ladybug for different tasks. At the end I think I only used Octopus for the evolutinary optimization of the building form, Horster for the HUD and a self-scripted permutation solver for the apartment optimization (There is no evolutionary solver for permutations yet). However the biggest help was the forum! There are so many answered questions already, that it is quite easy to deal with most problems and it really helps to understand GH! Thanks community and thanks David!

About

Translate

Search

Videos

  • Add Videos
  • View All

© 2024   Created by Scott Davidson.   Powered by

Badges  |  Report an Issue  |  Terms of Service