Modeling Vertical Migrations of Zooplankton Based on Maximizing Fitness

The purpose of the work is to calculate the evolutionarily stable strategy of zooplankton diel vertical migrations from known data of the environment using principles of evolutionary optimality and selection. At the first stage of the research, the fitness function is identified using artificial neural network technologies. The training sample is formed based on empirical observations. It includes pairwise comparison results of the selective advantages of a certain set of species. Key parameters of each strategy are calculated: energy gain from ingested food, metabolic losses, energy costs on movement, population losses from predation and unfavorable living conditions. The problem of finding coefficients of the fitness function is reduced to a classification problem. The single-layer neural network is built to solve this problem. The use of this technology allows one to construct the fitness function in the form of a linear convolution of key parameters with identified coefficients. At the second stage, an evolutionarily stable strategy of the zooplankton behavior is found by maximizing the identified fitness function. The maximization problem is solved using optimal control methods. A feature of this work is the use of piecewise linear approximations of environmental factors: the distribution of food and predator depending on the depth. As a result of the study, mathematical and software tools have been created for modeling and analyzing the hereditary behavior of living organisms in an aquatic ecosystem. Mathematical modeling of diel vertical migrations of zooplankton in Saanich Bay has been carried out.