Navigating Design Space
April 29, 2012 in Compuational Design by Sivam
Understanding design as an exploration in design space. The key differences between optimization and exploration is also explained.
Feel free to comment. You can download this document if you wish.
Sivam,
Thanks for the lecture notes. i have few queries as below:
1) Under the sub-heading ‘Navigating the developmental space’, you have mentioned that “We need to visualise…Well before it gets to the parameterizable stage”. Could you please expplain that further, may be with an example. Here i dont quite understand what you mean by the term ‘Parameterizable stage’
2) ‘Difference b/w Parametric Model & Genetic Model’ – 2nd last and lest sentence “They also need to help…into recognizable species of design. In short…facilitate the human-driven generative design process”. Here are you trying to relate to “Capturing the commanality” in chapter 4.0?
3) in the current lecture notes, you have often mentioned about design stages namely early design stage, developmental stages, later and final stages. Are these stages the same stages that are involved in manual design process? or do computational stages are different to general manual design stages? What i want to know, is what in you opinion is the defination of early stage (sketch phase for manual design process) and developmental stage (space planning, etc in manual design process)?
4) Optimisation Vs Exploration – Cant we use both process in a single design? for instance we could assign a value to solar shading device in a design so we use optimisation method for that, but we cant assign value to experential aesthetics, for which we perform exploration method. This way we can concentrate only on few areas of design for exploration and let the computers optimise all thechnical bits?
5) Design Process – When i sit to design, say for example a comminity hall. After all site analysis, context study, and everything involved in pre-design stage, i kind of have a vague idea about the building and building envelope that might fit the design problem. So, in short, I most often than not pre-empt the design in my head during early stages and than fine tune it as we advance further. Is this the process followed while adopting CD? I mean, even to build a schematic Genetic Model, we should have an idea of what we are after isnt it? So that initial brief, parameters and designer’s GUT FEELING, is important even while adopting CD methodology…what are your thoughts on it?
aB
Discussion # 5 is covered in in chapter 6 and even some part of 4.
1) Now we can technically use design spaces only when we have a CAD model and parameters to drive it. But the concept of design spaces can be used even with out. So I am pushing you here to imagine sketches as being instances in designs space.
2) When you are exploring 1000s of designs you can easily get lost. A family tree will help you to understand where you are within design space.
3) Currently computational design only happens at very late stage – when the design is represented in parametric form. This is a serious limitation. We need to figure out how to use computational methods well before that. That would include layout as well.
4) Yes, you can use optimization and exploration in parallel. Optimization can help you generate knowledge about the performance space and is useful in that context.