Advances in cloud computing promise to radically enhance the development of built assets by automating certain design processes that have long been intuition-based.

According to Philip Bernstein, VP Strategic Industry Relations, Autodesk, enhanced access to staggering amounts of computational power will result in profound changes to the design and development of built assets around the world.

“We’re going to see some fundamental changes in the way that assets are built, designed and operated, and chief amongst them is what I call ‘anecdotal predictiveness,’” he said.

Bernstein believes the vast amounts of computational power accessible at low cost via the cloud means that certain decisions that were once made by experts based on years of experience and fine-honed pattern recognition could soon be automated.

“We now have access to infinite computing and infinite storage thanks to the cloud,” he said. “You want a million CPU cycles from Amazon, just get out your credit card – we can do it on my laptop right now. CPU cycles are no longer a precious resource and storage is practically free.”

The example of a comparatively simple infrastructure asset such as a highway can be highly instructive for understanding how computing power will change design-related decision-making.

“If you’re laying out a highway in the middle of the countryside, you want to figure out what’s the minimum amount of cut and fill to do the highway, in order move as little dirt as possible,” Bernstein said. “Any dirt that you move when you make a slice through a hill, you want to use that dirt to prop up the highway someplace else – ideally all the cut and fill perfectly balances so you don’t have to haul away or bring in any extra dirt.

“Doesn’t sound like much, but when you’re talking about a 15-mile highway it’s a mountain of dirt, and very expensive, so the processes for optimizing that are extremely well understood.”

Bernstein notes that while the challenge of minimizing cut and fill in road construction has already been mastered, highway design remains a complex process involving multiple other variables, with certain key decisions still dependent upon the intuitive assessments of veteran experts.

“What’s not so well understood is the relationship between optimizing the cut and fill approach and optimizing the construction sequencing itself – what’s the right strategy for building roads while using the minimum amount of time, as well as how to deal with other deeply interconnected variables, such as construction safety and cost,” Bernstein said.

“What happens today is an experienced highway engineer who’s built a lot of highways begins to build up an intuition about what works and what doesn’t, and she uses her intuition to make those decisions – I’ve seen this a million times before, and that’s what happening now in this current state.”

This state of affairs is on the verge of radical transformation, however, as a result of the cheap availability of huge amounts of computational power thanks to advances in cloud computing.

“We will be able to model those things and predict them with a great deal of accuracy in the foreseeable future,” Bernstein said. “So you’ll be able to actually interconnect those variables, and begin generating design solutions that represent a chain of those interconnected variables, and generate them, as opposed to create them out of your head.”

According to Bernstein, this revolution in computing power will enable architects and engineers to produce a huge host of design options based on set criteria before whittling them away to obtain an optimized solution.

“You can say to your computer – and highways are a really easy example because there are no considerations relating to beauty and proportion and other ineffable things – I want to get from here to here, so generate five thousand alternatives and show me what the trade offs are between cut and fill safety, highway safety, safety in construction, costs, and speed,” he said.

“And it’ll come up and say, it looks like we have 5,000 alternatives here, 3,465 fall outside the parameters of cost, so let’s throw those out. I have 1,500 alternatives left, well these 300, statistically we know from our construction kit strategy, we’re going kill at least three guys, so those options are out.

“And you’ve created a problem space, and you’re narrowing things based on a set of relationships which computationally now is very easy, when it used to be impossible.

“Now it’s relatively straightforward.”

This vastly enhanced computational power will be applicable to numerous other areas of design in relation to the built environment, including high-rise buildings and maximizing the amount of available light in an enclosed area.

Bernstein notes that this computational prowess will not lead to the replacement of human beings as the primary actors in the design of built assets, but will instead enhance their decision-making ability by simply automating those aspects of the process that should be subject to rigorous and inflexible criteria.

“There is a whole series of performance questions about built assets that shouldn’t be subject to judgement – you should just measure them,” he said. “Either it works or it doesn’t work, and what your judgement starts to become about is the interplay and the relationship between the pros and the trade-offs, because it doesn’t always have perfect equations.”

The ability of computers to effectively and impartially access certain measurable aspects of built asset performance will also improve as this practice because more widespread, due to the accumulation of more extensive knowledge and data on prior results.

“There’s going to be a ton of available data that didn’t used to be available – about what worked, what didn’t, which solutions did we try, how did they perform,” Bernstein said. “Figuring out what the traffic flows were like for highways and how much money we’re going to generate, how long the concrete lasts, how long it took the roadway to drain, how long it took us to repave it – all that data starts to aggregate.

“And you can go back and study it, and it will feed back into your next design loop.”