Generative Design

It is refreshing to see promising  research and useful methods emerging from lesser known quarters. Despite decades of  academic research the “layout problem” as it is called, is till today solved by intelligent guess work.

Christian Derix, director of Aedas R&D Computation Design Research (CDR) group seem to be close to cracking a rather long standing problem. A problem that obsessed early design researchers from the end of world war 2. Architects returning from the war seem to have been keen to shake out the “irrational: image of their professions. Their engineering colleagues got back to peaceful production and were focusing their efforts on improving production and making it efficient. 50~30% of production costs are attributed to what is called transport cost, or the cost of moving material from one place to the other.

This problem has been researched over and over again – till the cows come home. If you search journal articles on the layout problem  you will find a great abundance of them with 100s of different methods. You will be even more surprised with the range of applications ; from laying out factories, container yards, chips on circuit boards and even for arranging cows for efficient milking. These methods seem to work generally well if the problem is focused on optimizing a single dominant criteria, which is  often the minimization of transport cost.

These methods can theoretically be applied on multi-criteria problems – but only theoretically. Only if you know for sure what those criteria are. Only if you can measure them accurately. Only if you can then balance them all with a magic formula – which in an immensely engaging theoretical exercise in itself. But that does not seem to dissuade researchers. Because the application of their methods is left for lessor beings who remain adamant for decades not to use them. But if you wish to for some reason publish that one more paper on ” multi-criteria optimization of the layout problem ” -  there is a way around it. All you got to do, as many research do; start with a lamentation in your introduction “ that despite the many decades of research, and many methods proposed by researchers, it is well known that these methods are not used by industry ” . You are then excused to propose yet another method with the same fate.

One would imagine that in  engineering environments where measurable conditions of efficiency and cost reign, that these methods will be used – apparently not. Management, maintenance and many other mundane  issues seem to throw in the spanner making factory layout the cherished skill of a few experts.

A designerly approach

A better approach is accept the nature of design problems – with some honesty; that we really do not know the criteria we use, let along measure them. But we can jolly well recognize a potential solution when we see one emerging of out of our scribbles. Let us accept that we keep changing our mind during design. We should – because we develop a better and better understanding of the design problem that the related issues as we develop our designs.

Another researcher Mark Syp  is working on an ” advancement of the bubble diagram ” and is developing to a parametric conceptual tool to quickly organize and understand complex architectural programs in three dimensions.

So the answer lies  not in eliminating the human designer, but in assisting the designer by managing the constrains and the requirements that evolve through out the design process. Physical simulations engines are good at this. Many of them are  now seamlessly integrated in to most CAD packages. Generative design too can play a role here to create some initial configurations which the designer can further manipulate and refine. I have experimented with this approach and works superbly. This is certainly the way – that conceptual layouts will be solved in the future.

It seems now that we are closer to cracking the long standing layout problem. The conceptual stage of architectural design will benefit greatly from it,  thanks to the inspiring efforts of those like Christian Derix and Mark Syp.