Determination of recipes for manufacturing semiconductor devices

This invention describes a system that helps create new manufacturing instructions, called recipes, for making semiconductor devices. It works by first running real-world experiments and virtual computer simulations of the manufacturing process. The results from these experiments and simulations are then fed into a machine learning model, which learns how different parameters affect the final product. Once trained, the model can take desired specifications for a semiconductor device and automatically generate a precise new recipe to achieve those specifications.