Today we released our AI-based 1D and 1.5D (for miter cuts) nesting feature into cmExe! Starting from the requirements of our customers and working backward, this result is the culmination of a lot of blood sweat and tears coming from our product design, our data science / R&D, and our software development teams. This feature set allows users to perform the classic nesting operations called different things by different people: 1D cutting optimization, stock length optimization, 1D bar nesting. We allow the user to play with different parameters and run the optimizer numerous times to study and settle on the best possible output.
After honing in on the final result, the user can then feed this optimized cut list into our digital traveler in the “job order” tab for tracking of the fabrication process. Early feedback from users shows a savings of over 80 labour hours of manual entry per production release over the incumbent legacy nesting solution by allowing enhanced features in the user interface to group and batch process parts for nesting. In these early tests versus the incumbent, we are still studying the detailed data and results from our first few tests with customer data, but our 1D nesting results have yielded less waste which can add up to significant material cost savings when procuring bars for a 60 story tower facade.
This 1D nesting solution is available for all customers to use immediately in the cmExe product as well as outside the product via API call. If you have an interest in our API solver, please contact us at firstname.lastname@example.org.
When looking at this challenge and how we would solve this in our cmExe product, we chose to look at an old problem through a new lens. 1D nesting/optimization is an old problem, with many old solutions existing in various forms in the market. Our data science team analyzed existing typical algorithms that have been traditionally used and decided to start from a clean sheet of paper when looking at the most optimized way to solve these old 1D and 1.5D nesting problems.
cmExe - 1D nesting to digital part traveler workflow
Warning! This is about to get pretty technical, feel free to duck out now or, if you’re nerds like us, please read on for the nitty-gritty details.
Our solution leverages a framework called Genetic Algorithms in machine learning. A quick technical background, the best description of genetic algorithms that we found was from Jonathan Shapiro, part of the lecture notes in computer science book series “LNCS, volume 2049): “Genetic algorithms are stochastic search algorithms which act on a population of possible solutions. They are loosely based on the mechanics of population genetics and selection. The potential solutions are encoded as ‘genes’ – strings of characters from some alphabet. New solutions can be produced by ‘mutating’ members of the current population, and by ‘mating’ two solutions together to form a new solution. The better solutions are selected to breed and mutate and the worst ones are discarded…Genetic algorithms are used in artificial intelligence like other search algorithms are used in artificial intelligence – to search a space of potential solutions to find one which solves the problem”.
We took the framework of genetic algorithms and created a custom solution to solve this very specific problem for 1D nesting. The final 1D nesting solution in cmExe is extra powerful because it doesn’t need a pre-trained data set or machine learning, our genetic algorithm solution can generate the best possible result based on the user input without prior training needed. We simply generate random solutions from the input data, evaluate them, select the best ones, generate more solutions from them (mating/mutating) and continue with this process until we get the optimized result.
We would love to hear your feedback, send us your comments today! Much more to come!