A Preliminary Study on the Shelterbelt Structure in "Weather-windbreak-yield" System by Markov Decision Process

In "weather-windbreak-yield" system, the structure characteristics of the farmland shelterbelts are of importance for increasing the yields of crops. On the basis of the inter-disciplinary principle in relation to ecological dynamics, ecological-cybernetic principle, meta-synthetic methodology and Markov processes, mathematical modelling method, computational process and decision principle about the application of Markov decision process (MDP) are used to determine the structure of the farmland shelterbelts in relation to weather conditions. By means of the data obtained from Haizhou, Kangping County, the transition probabilities of the related states about the effects of the farmland shelterbelts are analyzed, which are of different structures on the meteorological yield departure of the maize crop on the basis of the fundamental principle of probability theory. The concepts of the two states (i=1 and i=2), related meteorological outputs and the action measures-"ventilating structure shelterbelt"(a=b) and "open-out-structure shelterbelt"(a=c) are given. Then, the optimal policy and the improvement optimal policy from the computational results suggests based on the discount model of MDP with the discount rate of 0.9. The results show that the optimal policy is to choose the ecological-cybernetic action open-out-structure shelterbelt with ventilating coefficient of 0.35 in the given complex ecological-cybernetic system which is formed under the conditions of the local weather-climate and the agricultural-and forest-crop states.