| With the development of science and technology, the emergence and development of intelligent systems provide strong support to modern enterprises and institutions. They change the traditional mode of work, greatly improved work efficiency. Urban food emergency command system as a typical intelligent application of the field of food security, Its establishment and improvement could make effective supervision and scientific analysis and a reasonable forecast and strong control to food production, processing, transportation, market supply and marketing. What is more, it can timely feedback all kinds of food the field of information to the food department relevant businesses in charge, the relevant departments can be able to quickly take the appropriate measures to optimize the allocation of resources, reduce losses, reduce losses and risk, maintain social stability.In urban food emergency command system, early warning model as well as early warning plans are the core part of the system. They play a key role in the early warning function. Through food data collection, analysis, combination, processing and finishing, early warning model can make accurate predictions for scientific warning. When the warning level produced by early warning model, the system set off the early warning plan corresponding to the early warning level. The implementation plan process as soon as possible the implementation of preventive measures. Currently, there are lots of modeling programs. They have advantages and disadvantages of each model created in the early warning accuracy, complexity, etc. However, due to the grain field warning model with large dimensions and the characteristics of nonlinear and small sample size and unbalanced, their methods cannot meet well all the requirements. In addition, the traditional early warning plan that customized solutions with fixed mode complex, customized, long development cycle, is unable to meet needs that are simple, flexible, and fast to customize plan in the system.This article made the key research work by the following:(1) Through the demand analysis for early warning model in real environment, and pointing against fuzzy comprehensive evaluation model having no self-learning ability and poor adaptability, the paper proposes a method that constructed early warning model based on fuzzy neural network combining the complementary fuzzy evaluation with neural network modeling. In the case of ambiguity with good judgment, the model enhances fuzzy evaluation method self-learning ability. The early-warning model has good adaptability and accuracy of the forecasts. It can improve early warning accuracy rate, reduce the complexity of the algorithm. The validity of the fuzzy neural network learning algorithm was verified by the experiment. (2) Aimed to the low reuse, poor portability and operability in traditional early warning plans. The paper proposes a plan customized solutions based on JBPM (Java Business Process Management). Workflow technology has advantages in flexibility, simplicity and short development cycle. It can meet demands of early warning plans customized fast and flexibly. Good effects are verified by the experiment in new early warning plan.(3) New early warning model and early warning plans were applied in the development of Nanning food emergency command system. When a warning level is produced in the early warning model, the system starts the corresponding warning level plans. Early warning plan could transfer in the order pre-designed, the function of each node will be completed when the plan transfer to the node. It is beneficial in taking effective measures to early warning and mitigation rescue. Combined with the relevant Grain Bureau data, tested by the experimental, the model receives rate of100%in warning accuracy. New model and plans meet the user’s needs basically. They provide a theoretical basis on food security early warning, supervision and decision making. |
PDF Full Text Request