Mackerel and Herring on the Move A Migration Research Based on Deep Q-network

Abstract

With the development of computer technology, reinforcement learning algorithm is gradually applied in ecology. When the temperature changes greatly, mackerel and herring will not be able to sustain their lives, at which time they will choose to migrate to habitat suitable for their survival and reproduction. This will greatly affect the economic benefits of fishing companies which fishing in the fixed sea area. Therefore, this paper tries to find a reasonable method to predict the migration location of these two species of fish in the foreseeable future. To understand more clearly the survival status of herring and mackerel in the waters near Scotland, we collected a large amount of reliable data and established a prediction model based on Deep Q-Network (DQN). Given that there will be no dangerous situation at sea to affect the fishing operation, the habitats of herring and mackerel will gradually move to high latitudes in the next 50 years.