Tilapia feeding decision system based on adaptive neuro-fuzzy inference

Abstract

In industrial recirculating aquaculture, the feed required by fish accounts for a major part of the total expenditure. In this paper, a multi-factor decision making system based on aggregation FIFFB of fish feeding behavior, water temperature T of environmental factors and biomass weight W was proposed to solve the problem of feed waste under the traditional mode. To verify the performance of this model, a fuzzy inference FIS model is constructed for comparison. The experimental results show that the root mean square error (RMSE) and mean absolute error (MAE) between the predicted and actual feeding amount of ANFIS triplet are 0.78 and 0.19, respectively, which are much lower than the FIS model, and this model is more suitable for predicting the feeding amount decision. At the same time, growth parameters such as WGR, FMAE, K and FCR were compared. The fish growth specifications were fatter and the economic benefits were higher, and the feed conversion rate was increased by 12.35%. Therefore, the triplet ANFIS feeding prediction and decision-making system based on fish aggregation degree, water temperature and body weight is effective and has guiding significance for the precise feeding design of unmanned aquaculture.