Improvement And Application Of Particle Swarm Grey Markov Chain

China is a big country in the cultivation, production and consumption of dates. Dates industry is an important part of agricultural production in China. Therefore, scientific accurately predict annual output of Chinese jujube in the agricultural sector formulated relevant policies and reasonable production planning has an important reference value. At present, relating grain yield prediction is rich in the aspects of the theory and methods, but specific to the red dates, the yield prediction of research is less, so this paper established three reasonable mathematical model: grey GM(1,1) model, grey Markov chain model and based on Particle Swarm Optimization of the improved grey Markov chain model and prediction of annual output of Chinese jujube respectively, and the prediction precision is analysed.Grey system theory is a method to study the problems of small sample, poor information and uncertainty, it has been widely used in many fields such as industry, agriculture, society, environment, ecology and so on, Among them, the most basic and applied model is the grey GM(1,1) model. Due to the natural environment, social environment and agricultural policy, and many other factors, The annual yield of Chinese dates has randomness and volatility, it can be regarded as a grey system, so, in this paper, we first consider the establishment of gray GM(1,1) model to obtain the basic trend of red dates in recent years.Traditional GM(1,1) model can reveal the development trend, but intends to the random fluctuations of the larger data of poor degree, prediction accuracy is not high; Markov chain system based on the current status and changing trend to predict system could the state. It is used on the machine volatility larger data sequence with, in view of this, this paper further uses Markov chain of years the jujube yield grey GM(1,1) model is modified, to establish the grey Markov chain model, make up the GM(1,1) model of non monotonic sequence prediction effect is poor.Particle swarm algorithm is a bionic cluster optimization algorithm, through the simulation of bird foraging process to search for the optimal solution. It needs less adjustable parameters, has a strong global search capability, is applied to various optimization problems. In order to further improve the accuracy of prediction model, this paper uses particle swarm algorithm to optimize the red dates annual output of Grey Markov chain get the model, particle swarm optimization and grey Markov chain model. Different from the constant practice of inertia factor standard particle swarm algorithm, the model with a linear decreasing inertia factor by way of optimizing, to enhance the convergence of the algorithm, put forward a new fitness function, to avoid the tedious calculation error square general particle swarm algorithm and, reduce the amount of calculation.This paper chooses China’s National Bureau of statistics dates from 2005 to 2012 annual output data as the research object, the establishment of GM(1,1) model and gray Markov chain model and particle swarm the grey Markov chain model. Using these three models to predict the annual yield of Chinese dates in 2006-2012, and compared with the real data, the average relative error of the three models were 3.60%, 1.15%, 0.67%. The results show that the annual output of red dates particle swarm the grey Markov chain model has higher prediction accuracy and to date production enterprises and agriculture departments formulate date production planning and policy to provide data to support. Finally, using particle swarm optimization and grey Markov chain model forecast to 2013- 2019 annual output of Chinese dates.