The Optimized Model Based On Grey Markov Chain And Its Application In Prediction Of Tea Output

Tea industry is an important component of the national economy, especially an integral part of agricultural production. Thus, tea industry is of great Significance, and can not be ignored. The smooth development of tea industry is important for the stability of social life, the modernization as well as the building of a harmonious society. To predict the tea output in China will help the relevant departments quantitatively analyze the current economic situation, scientifically predict, and make the reasonably practicable policy, so as to promote China's tea industry to develop reasonably and orderly.So far, many scholars have done a lot of work on the prediction of grain yield, and the theories of prediction have made significant progress. However, the research work to predict tea output has been done before is little and not enough; there still is lots of work for us to do. From a variety of the research data, we find that the mean to prediction the tea output is only one single predicting model, such as the grey prediction model, and a lot of good prediction model has not been applied and developed. Therefore, the paper tries to present better theoretical research and empirical application on the tea production predicting.In this paper, Markov chain and the gray model are combined to establish a mixed grey Markov prediction model to predict the output of tea. Grey theory is a predicting method for the grey system, including non deterministic factors. Based on the past or present known or undetermined information, which is the so-called gray information, it is principle to establish a gray model (Grey Model, referred to as GM) to predict the unknown information. GM(1,1) is the basically and relatively simple prediction model. Though the traditional GM(1,1) has a poor effect to fit the random data sequence and predict inaccurately, it can well forecast the general developing and changing trend of the data series. Markov chain can predict the possible state in the next stage based on the present state and the trends of the system variables, and thus provides a basis for the decision-making. Markov process can determine the status of the transfer rule, well be applied to predict some problem with the system's change having random fluctuations. As a result of natural and social conditions, tea production has a lot of randomness and uncertainty. Therefore, the formation of the tea output can be regarded as a dynamic grey system, we establish GM(1,1) prediction model. As the tea output is volatile, and Markov chain has some advantages on predicting the random fluctuation, we use Markov state transition matrix to correct the grey prediction model, by which the Grey Markov model is established. The two approaches can complement each other to improve accuracy of the prediction model. To further improve the prediction accuracy of the model, this paper presents a particle swarm algorithm (PSO) to optimize the Grey Markov model, establishes a new forecasting model, called the PSO-Grey Markov Model. Particle swarm algorithm is a simulation of the evolution method of birds seeking food, which adaptively adjust search direction based on the principle of shared information, mutual coordination and competition between individuals of the population. The algorithm has the advantages of simplicity, less parameters needed to be adjusted, having nothing to do with the characteristic information of the problem, strong global search ability and so on. The grey Markov model makes good use of the good characteristics of particle swarm algorithm to further improve the prediction accuracy.This paper applies the presented theory above to the prediction of China's tea output, and gets a satisfactory result. The tests show that the theory can be applied in the prediction of China's tea output, as well as afford a theoretical basis and reference for China's tea production planning and decision-making.