Feeding requirements of sea bream (Sparus aurata) larvae under intensive rearing conditions: Development of a fuzzy logic controller for feeding

Abstract Applied methods for larval rearing suffer from the difficulty of accurately predicting larvae quantitative feeding requirements for optimum growth. Even when known, requirements change according to the behavior of each population reared. Computer based feeding systems require tools for adapting the delivered amounts of food to the needs of the reared population. In this work the daily feeding requirements of sea bream ( Sparus aurata ) larvae were estimated, and based on these estimations a fuzzy logic controller (FLC) for feeding was designed and developed. Sea bream larvae were reared, in triplicate, under intensive conditions in a pilot scale rearing system for 20 days with the pseudo-green water method. Larvae reached 6.7±0.6 mm (mean value and S.D.) at the end of the experimental period while the mean survival of the populations was 48.0±12.6%. Consumption was 0.017±0.009 mg of carbon per individual at first feeding and increased to 0.198±0.032 mg at day 20 post-hatching. The findings were used for the development of a FLC for feeding that adapts the amount of food delivered to the larvae according to changes in the reared population. The FLC utilizes five linguistic variables describing the state of the population and a rule base composed of 316 rules. The developed FLC allows the use of computer based feeding systems during rearing of early larval stages.