The AIGODS project has opened a new methodological line with the beginning of its Genetic Programming tests, marking an important step in the optimisation and benchmarking strand of the project. In a research framework already grounded in deep learning for yield estimation, this expansion reflects a deliberate effort to explore whether alternative artificial intelligence paradigms can also contribute meaningfully to modelling and comparative assessment in the Douro Demarcated Region (DDR).
Within AIGODS, this new direction has been introduced not as a replacement for the project’s core LSTM-based forecasting workflow, but as a complementary line aimed at widening the methodological scope of the final comparison. Once the main yield estimation branch had already established a stable reference baseline, it became scientifically relevant to assess whether a different modelling logic could also be tested against the same regional problem under a controlled comparative framework. This is the context in which the project has now begun its first SLIM-GSGP optimisation tests.
As Leonardo Vanneschi notes,
“Exploring Genetic Programming within AIGODS allows us to test a different modelling logic against the same regional problem, strengthening the project’s comparative framework and opening new possibilities for future optimisation.”
This perspective reflects the role of Genetic Programming within the project at this stage, not as a definitive modelling shift, but as a technically meaningful extension of the benchmarking space already created by the forecasting branch.
While these first experiments represent a relevant advance in the project’s analytical depth, this research line is still under consolidation and should be interpreted with appropriate caution. Even so, the beginning of the GP strand confirms that AIGODS continues to evolve with methodological rigour, testing new avenues of optimisation in a way that remains both scientifically grounded and potentially valuable for future comparative work. Further technical details will be shared in due course.
Login Required
Loading login form...