Dynamic migratory beekeeping route recommendation based on spatio-temporal distribution of nectar sources

Due to the lack of scientific guidance, most migratory beekeepers currently arrange their migratory beekeeping routes by experience. The production mode is extensive, and the quality and efficiency need to be improved. Therefore, this study investigates a dynamic migratory route recommendation problem considering the stochastic yield of nectar sources and uncertain disastrous weather events, by which the beekeeper can dynamically follow the practical and effective recommended route to cope with production risk to reach a better revenue. To this end, the problem is first formulated as a Markov decision process considering various flowering durations of nectar sources, migration costs and time, the prices of bee products, and, most importantly, uncertain yields. Then, an approximate dynamic programming algorithm incorporated with offline and online learning procedures is proposed to deal with the curse of dimensionality. Several acceleration methods are also provided to solve the problem more efficiently. The conducted numerical study shows that the proposed model and algorithm perform well in approximation precise and computational efficiency. Finally, the computational results show that the proposed migratory beekeeping route recommendation method effectively deals with yield uncertainty and significantly improves the migratory beekeeping revenue.