Optimal fish densities and farm locations in Norwegian fjords: a framework to use a PSO algorithm to optimize an agent-based model to simulate fish disease dynamics

Abstract

Risk of pathogen transmission in Norwegian fjords depends on two main factors: location of farms and the density of fish in each farm. This paper presents a novel method to find the optimal values of these two variables that yield the optimal aquaculture system with a minimum risk of spreading disease and high fish production. For this purpose, agent-based models (ABMs) are used to simulate and analyze fish disease dynamics within and between fish farms in Norwegian fjords. Moreover, a modified particle swarm optimization (PSO) algorithm is used to identify the optimal values of fish density and farm’s location for each farm. The objective function is defined as being the weighted sum between the fish density and the infection risk. We validated the PSO algorithm with the optimal objective function by demonstrating the capability of the algorithm to drive the system to produce an expected behavior and output in tested, known scenarios. The simulation results demonstrate the ability of the PSO algorithm to converge rapidly to the optimal solution. In only 18 iterations, it finds an optimal solution that is three times larger than the initial fish farm density and in a location that keeps the risk of infection at an accepted level. The use of the PSO algorithm in finding optimal parameter values of ABMs will open for new applications of the model in aquaculture industry management, such as planning for a sustainable aquaculture industry.