As a construction material, timber is prized for its strength, availability, and sustainability. And yet, planning and designing a floor plan for a timber structure is no easy feat, thanks to the thousands of possible combinations the material can be placed even for the simplest floor shapes. To solve this challenge, one UK company is borrowing some cues from the science of genetic programming.
‘Artificially Intelligent Designers’ Solve a Longstanding Problem
Even digital modeling of such floor plan designs using CAD or similar software programs requires a great deal of human intervention and time, and optimal results aren’t guaranteed. This was a recurring problem for stair and floor solutions expert Staircraft, so they developed their own software. The solution helps narrow down the number of combinations and designs. According to Forbes, the company’s software solution helped a little but still didn’t cover all combinations, potentially leading to suboptimal floor plan designs.
To optimize their design process even further, Staircraft partnered with custom artificial intelligence solutions provider Brainpool.ai to devise an AI system capable of running through those thousands of permutations to find the optimal design for a timber floor. To achieve this, the team enlisted the help of PhD-level data engineers, AI experts, and mathematicians, who leveraged their expertise of the latest research in genetic programming to come up with a brand new solution.
“Designing a floor plan involves processing a large number of combinations leading to hundreds of thousands of possible designs. Using the latest advances of Evolutionary Search, we can study all these potential combinations and eventually create a true optimum solution applicable to different scenarios for every client,” Brainpool.ai co-founder and CEO Kasia Borowska writes in Forbes.
“AI allows us to cover all geometrical shapes more rapidly and efficiently, getting an optimum solution in a fully automated manner. Hence, this method creates an ‘artificially intelligent designer’ that is much more efficient than a human designer.”
Genetic programming, Borowska writes, uses the principles of evolution and natural selection for optimization. Designs are selected and created at random and run through the system, which tests them for factors like cost and structural feasibility. Each design is then given a “fitness” value, and each subsequent generation of random designs created includes the optimized features from the previous attempts. This process is then repeated, with each new iteration incorporating the previous generation’s best features until the optimum design is found. The method stems down from how the best human genetic traits tend to be the ones passed down from generation to generation.
How Optimizing Design Reduces Waste, Curbs Environmental Impact
The construction industry generates hundreds of millions of tons of waste each year. According to the Environmental Protection Agency (EPA), more than 600 million tons of construction and demolition debris were generated in 2018 alone. Optimizing designs using genetic programming techniques can help reduce inefficiencies and excessive waste created by scrapped designs that failed to live up to expectations.
Borowska delves into some of the efficiency-boosting and waste-reducing benefits of the automated design system.
- Less Material Waste. Borowska writes the average British home contains approximately 100 linear meters of joists, with a combined weight of about 300 kg (660 lbs). She says AI-optimized design saves around 8% of those materials, or approximately 25 kg (55 lbs) per home constructed. With some 160,000 homes built in the UK every year that translates to more than 4,000 tons of wood waste saved each year.
- Reduced Transportation Costs and Pollution. Even that 8% materials reduction will go a long way towards reducing a construction company’s carbon footprint, Borowska writes. The savings on materials alone will eliminate the need for more than 2,500 truck deliveries each year.
- Improving the Efficiency of the Construction Workforce. Borowska writes that employing an AI system can slash the time spent on design by as much as 50%.
Borrowing from genetic programming seems a brilliant solution to a common construction problem. The hope is that this same method can be applied to other areas of construction to improve building efficiency, reduce waste, and devise the optimal design for any project more quickly.