Aquaculture Reports publication

SAFE partner CIIMAR has just published the open access publication “Predicting weight dispersion in seabass aquaculture using Discrete Event System simulation and Machine Learning modeling“.  It presents an innovative approach in predicting fish weight dispersion in farmed European seabass (Dicentrarchus labrax), one of the two major fish species cultivated throughout the Mediterranean. During its fattening cycle, several grading operations are carried out to minimize growth dispersion and this method represents a significant advance in the planning and management of seabass aquaculture.

The process starts with a Discrete Event System (DES) model simulating the feed-fish-water dynamics, accounting for individual fish metabolism. A surrogate machine learning (ML) regressor model, that has been trained on the DES data is then used to predict the growth distribution. The model can be calibrated and customized for specific fish stocks and production tanks. The preliminary results from two trials with European seabass (D. labrax) in 21 tanks showed the efficiently of the method. The results from the simulation models achieved a R2 of 99.9 % and a Mean Absolute Percentage Error (MAPE) of 1.1 % for the prediction of mean final weight and a R2 of 90.3 % with a MAPE of 8.1 % for the standard deviation of final weight. With the few effective prediction tools in the aquaculture industry, the proposed methodology has the potential to reduce inefficiencies and risks while increasing sustainability and profitability by optimizing fish farming practices.

The publication will be presented in the upcoming Aquaculture Reports, issue 38, which will be released in October 2024 and is available here: https://www.sciencedirect.com/science/article/pii/S2352513424004034

Subscribe to our newsletter

Latest news items