A new feature in the Alchemite™ Suite machine learning software makes it faster and easier to include cost analysis when optimizing formulations, materials, and other products and processes. With cost factors always front-of-mind in sectors such as chemicals, materials, and FMCG, Alchemite™ provides guidance that reduces the number of experiments needed to uncover novel product solutions that can save dollars without sacrificing performance. This drives cost savings in R&D (through shorter experimental cycles), in production (by lowering resource requirements), and in the use-phase (through more efficient end-products).
The new feature works by making it quicker and easier to enter cost data when building a machine learning (ML) model of your system. It can even use the generative AI capabilities of Alchemite™ Insight to suggest cost values for key inputs. Building this cost information and the goal of minimizing cost into your ML model enables its use to propose new experimental designs that target lower-cost outcomes, balanced against other objectives such as physical or environmental properties. By following this adaptive design of experiments (DOE) approach, these goals are typically achieved with 50-80% fewer tests than when using conventional DOE methods. Alchemite™ also provides graphical analytics that deliver deep insights – for example, identifying which input factors are most significant in driving costs.
See the feature demonstrated in this video using the example of finding lower-cost steel compositions:
