New modelling of complex fish migration by application of chaos theory and neural network

Rules or patterns of movement or migration were still vague even for the main commercial fishes due to different routs in scale or in different, times resulting from complex environments to complex behaviour concept. The quantitative model of fish migration has been investigated using chaos theory to mimic more realistic fish movements by time steps from environmental and biological stimuli. The model uses three steps within a model neural network such as input stimuli, central decision‐making and response output in fish movements. The stimuli in the first step include the main physical (temperature, salinity, light, flow etc .) and biotic factors (prey, predator, life cycle etc .) which could be quantified as intensity parameters which were then normalized as ratios. The decision‐making process can be generated available signals for motor neuron using Lorenz chaos equations by the relevant stimuli. The response of fish movements from the output signal representing speed and direction can be re‐regulated as object‐oriented migration depending on physiological state or life cycle by third response filtering. The simulation results seen as 2‐dimensional seasonal migration for demersal fishes in the southern sea of the Korean Peninsula represented more realistic meandering tracks than the interpolated tracks in previous reports.