Development of the Aquafeed Algorithm for Multi-Objective Optimization of Aquaculture Feed Formulations Using Nutritional Composition, Growth Performance and Feed Cost Compared with Linear Programming and Genetic Algorithms

Abstract

Aquaculture feed formulation presents a complex multi-objective optimization problem requiring simultaneous consideration of nutritional adequacy, growth potential, and production costs. This study develops AquaFeed, a computational optimization algorithm integrating nutritional modelling, adaptive evolutionary search, constraint handling, and Pareto-based solution selection. The proposed framework incorporates ingredient composition, nutritional requirements, formulation costs, and predictive suitability measures within a mathematical optimization architecture. Reproducible Python experiments compare AquaFeed with conventional linear programming and genetic algorithms using synthetic ingredient datasets and repeated computational trials. Performance evaluation considers nutritional feasibility, economic efficiency, solution diversity, computational runtime, and statistical differences between optimization approaches. The computational findings reveal distinct trade-offs among the algorithms, with linear programming favoring minimum formulation costs, genetic algorithms achieving higher nutritional suitability, and AquaFeed providing alternative Pareto-based nutritional–economic compromises. The proposed algorithm demonstrates potential for supporting flexible ingredient selection and constrained feed formulation decisions. However, its practical effectiveness requires further evaluation using experimentally validated ingredient compositions, species-specific nutrient requirements, calibrated growth prediction models, and controlled aquaculture feeding trials to establish biological reliability, computational scalability, and applicability across production systems.