Process Cybernetics LLC (ProCyber for short) brings to industrial practice two recent Machine Learning algorithmic that generalize the traditional design of experiments (DoE) methodology.
Both generalizations aim to extend the power of DoE with respect to process dynamics. The first generalization, called the Design of Dynamic Experiments (DoDE), enables the design of experiments were some or all of the factors, representing operating conditions, can be functions of time. The second generalization, called Dynamic Response Surface Methodology (DRSM), enables the modeling of time-resolved data from a set of experiments. It enables us to effectively and efficiently capture the dynamic characteristics of a process in a data-driven model.
Design of Dynamic Experiments (DoDE)
The classical DoE methodology is not able to design experiments with time-varying conditions throughout the experiments. The Design of Dynamic Experiments (DoDE) methodology is a generalization of the classical DoE.
For details on how to design these experiments, see here
Dynamic Response Surface Methodology (DRSM) Models
In the classical DoE methodology, the models estimated from the data collected at the end of the experiments are called Response Surface Methodology (RSM) models. With the present-day availability of time-resolved data, one needs a type of model that efficiently represents the time dependence of the measured process outputs. This is achieved by the Dynamic Response Surface Methodology (DRSM), a generalization of the RSM modes.
For detailed information about the DRSM models, see here.
Knowledge-Driven Models
In cases where a detailed knowledge of the inner workings of a process is known, a Knowledge-Driven model should be postulated, and its unknown parameters estimated using experimental data. Such models are also called “fundamental” or “first-principles” models. For such a model to be accurate, one must worry whether all the estimated parameters are significant and whether the model represents all the non-random information in the data. Cross-validation of the model against new data is always a desirable task. Detailed information about these issues is given here.
Let's talk
We have collaborated with a series of big and medium-sized companies to benchmark these algorithmic technologies. We have applied them in bulk chemicals and specialty products as well as in pharmaceuticals and biopharmaceuticals. Likewise, we are ready to help your company benefit from these advances.
Arrange a presentation
