About
Ten years of optimisation algorithms behind Cutlist Evolution and SmartCut
Cutlist Evolution and SmartCut are built by the same team, and use the same proprietary optimisation engine. Cutlist Evolution is the web-based cutting-plan tool and SmartCut is the API. The e-commerce platform extends the optimisation API.
The core algorithms have had 10 years of development behind them and have run 1M+ optimisations. The key challenge when developing an optimisation engine is to balance efficiency and performance with ease of cutting and suitability for manual and computer-controlled saws.
Optimisation code is unusually easy to make worse by accident. A change that saves a sheet on one job can quietly cost one on a hundred others, and you will not notice from testing it by hand. So the engine has its own internal algorithm lab, and no change reaches a cutting list without going through it.
Every candidate is measured against the current engine over the same library of real production jobs, and it ships only if it is genuinely better across the board. It is slower than shipping on instinct, and it is the reason the numbers below hold up.
Cutting-optimisation software is unusually easy to make claims about and unusually hard to check. So we measure and thoroughly benchmark instead: 2,083 real production jobs in the benchmark library, run against competing tools on identical inputs, with the results published in full — including where we do not win. 82% of jobs finish on the fewest stock the parts could physically fit into.
The comparison and benchmark articles are written and measured by Jon Gibson, which is why they carry his byline rather than a company one. Where a competitor's capability is described it comes from their own published material, dated. Where a number is quoted, the method behind it is in the article.