Dear Francisco Maioli,

I would suggest you a mathematical on the Artificial Intelligence and the Idea of Digital Twins based Model for cattle/beef Production to develop.

You would then use the various Mathematical Tools (like Difference Equations, Dynamic Programming, Optimal Control, etc.) in the Digital Twin.

See Fig. 2 (Model of the Realization of Intelligent Services using Digital Twins) at the address:

https://www.researchgate.net/publication/366167747_Digital_Twins_in_IoT

Attached you have some Literature Sources to explain my Idea in more Detail.

Best regards

Anatol Badach

Benjamin Schleich, Nabil Anwer, Luc Mathieu, Sandro Wartzack: Shaping the digital twin for design and production engineering; CIRP Annals, Vol. 66, Issue 1, 2017, DOI: 10.1016/j.cirp.2017.04.040

Fei Tao, Fangyuan Sui, Ang Liu, Qinglin Qi, Meng Zhang, et al.: Digital twin-driven product design framework; International Journal of Production Research, Feb 2018; DOI: 10.1080/00207543.2018.1443229

Igor Halenar, Martin Juhas, Bohuslava Juhasova, Dmitrii Borkin:Virtualization of Production Using Digital Twin Technology; 20th International Carpathian Control Conference (ICCC), May 2019, DOI: 10.1109/CarpathianCC.2019.87659

Robert Miehe, Lara Waltersmann, Alexander Sauer, Thomas Bauernhansl: „Sustainable production and the role of digital twins – Basic reflections and perspectives“; Journal of Advanced Manufacturing and Processing, Vol. 3, Issue 2, Apr 2021, DOI: 10.1002/amp2.10078

T. Hiller, C. Cevirgen, P. Nyhuis: Exploring The Potential Of Digital Twins For Production Control &Monitoring; Journal of Production Systems and Logistics, Vol. 1, Mar 2021, DOI: 10.15488/10599

It might be a Markov decision process, or possibly a multi-stage stochastic program. Depends on the details.

I am not any idea about the topic

Depends on a number of factors and what you consider more suitable for your model. Do variables change in a continuous manner or at the end of each time period? In the former case, you could use a system of differential equations, in the latter case you may use difference equations, iterated maps or a Markov chain. This will only work if your agents are sufficiently homogeneous and you can model the process on the basis of a limited number of equations but given the number of factors and their impact on different producers you wish to model (also the dynamics are likely to be nonlinear), I suggest to create an agent-based model (NetLogo is fairly easy to learn).

Dear Francisco

I suggest using optimal control models.

Optimum Control is a powerful tool for your topic. If you decide to model with optimal control and you want help, I can help you.

good luck

if you are trying to apply your model to data (as opposed to a purely theoretical approach) then much depends on the nature of the data. However, while very discrete observations are most compatible with difference equations--how many depends on how you want to describe the phenomena. if you are not data oriented, or if the data permits, dynamic analysis is usually easier to handle with differential equations. the way you phrase the question seems to imply an attempt to determine whether responses to shocks are indeed optimal. If so try modeling an optimal control model using differential equations and solve for optimal responses, then compare actual responses to the ones the model finds would be optimal.