Natural ressources are becoming scarce, inflation has raised putting pressure on costs and carbon footprint is more than ever a concern for large industrials. Making vehicules and aircrafts consumes huge quantities of material (aluminum, steel, titanium, composites, etc).

D3S team has developed optimization algorithms to nest parts and massively reduce material consumption.


We work with x00’s Gb of engineering data (3D and BoM), purchase orders, jobcards and workorders. Expert manufacturing rules must be understood and modeled to optimize under constraints (clamping, allowances, fiber direction, table / machine dimensions, etc).

Our Tech Stack includes : Python, C++, PyTorch … and of course state-of-the-art librairies from our lab ;


Project objective was to reduce the consumption of material within the supply chain.

X,X b$ material are purchased each year covering aluminum, titanium, special alloys in several forms (plate, sheet, extrusion, forgings, …). Contracts are in place with material providers (Mills) enable to secure availability & prices, but the volume is spread within the full supply chain. This objective can only be reached with a data driven approach. D3S Data scientists developed algorithms to improve the precision of material forecast, optimize material usage with massive nesting capabilities, identify parts with the highest impact potential and control recycling value.


Full leverage of Analytics through next gen algorithms, massively optimizing patterns for mono-part or multi-parts nesting. Algorithms define pre-cutting and nesting scenarios with specific rules and constrains : saw cutting, water jet, milling, panoply, etc. They have been designed with a multi layers architecture combining a genetic core with heuristics.

Our customer achieved 20% to 30% material consumption reduction, leading both to high savings and a lower CO2 impact.

Algorithms are now embedded in our Digital Costing solution