Are You A Consequentialist?

…modeling the consequences of possible actions and select those actions that yield the best outcome? Consequentialism its the class of normative ethical theories (study of ethical actions) that could be simplified by the saying "the ends justify the means"

Adaptive control

I borrow this terminology for a class of pure technical innovations…innovations that want to achieve certain objectives. Say, quality features, a preferred behavior, a return at risk profile…of an outcome of a (production) process - a product, a good…

However you do it, by a mix of analytics or predictions you finally need to control your processes…you do it model based or data driven...its an inverse problem and consequently ill posed…

I've conducted quite a lot projects in this field: the best way to do it is by constant recalibration of control models …adaptive control…thats foundation is the constant estimation of changing parameters…a difficult task with a lot of traps.

We often think, physical modeling in contrast to, say, social or economical...is "real". But this is not always true. The prediction of a physical effect is not so easy as we often think. The theory of complex systems come into play…It matters what context we write into our models.

To solve a real problem, you often need to ride the waves of a real behavior and use tricky inversion techniques to recalibrate your models…

The Blank Swan of metal treatment

Let us take a numerically controlled metal treatment machine. Say, a forming machine. It "trades" a shape and the return pays in accuracy of that shape.  To implement the control,  you need to understand the framework of the elastoplasticity theory and the complexity and limits of its mathematics derived from the mathematics of continuum mechanics. 

Deformation is decomposed into elastic and plastic parts and for simplicity decompositions shall determine stress and kinematical quantities. The resulting PDE system can be solved by advanced numerical schemes (say, Finite Elements).

They "only" need to be calibrated related to the physical properties of the material that are dependent on recipes, and properties that are result of the process? And work as predictive models for final shapes, right? No Way.

The metal you buy change their specified properties during the treatment…headache begins.

So, you need to recalibrate your models during the forming process ("continuously"), say, by observing, say, force trajectories. The closer you come to the final shape the more your system knows about the concrete material (including its elastic "memory") and the better it can explain and predict...

Calibration and re-calalibration may need the application of clever machine learning techniques

The Blank Swan is Elie Ayache's original treatise of financial markets…it criticizes the naive treatment of predictability…

It may be surprising, but derivative pricing and metal treatment have much in common. Their adaptive control capabilities need multi-model and multi-method techniques and a lot of clever inverting techniques…

Being a consequentialist needs the full understanding of this framework.