The digital twin model is based on understanding the entire fish organism as an adaptive agent—a robust and testable biological theory that is implemented in computer code. This means that the model does not simply describe a specific aspect of nutrition, energetics, growth, or fish behavior in the form of an equation or a system of equations. Instead, the digital twin aims to function as a digital organism, simulating the most crucial aspects of physiology, neurobiology, and behavior in a digital environment.
The agent can therefore act autonomously, making decisions in response to both internal and external environments with continuous feedback at multiple biological levels. This means, for example, that the simulation system aims to predict voluntary behavior and feed intake of the fish. This nutritional digital twin enables the execution of various scenarios and predicts responses, including unforeseen, emergent, and stochastic effects. Such a capability provides an essential tool for decision support and operational optimization, and it can be run within the AI-controlled precision aquaculture environment of iBOSS, developed as part of the iFishienci project.