Research And Implementation On Parameter Identification Of Greenhouse Mechanism Model Based On LF-PPSO

With the development of agricultural technology,agricultural facilities has become an important symbol of agricultural modernization.Modern greenhouse crops able to provide a suitable environment for the growth of high quality and efficient crop production.Among them,the environmental regulation as the primary means to regulate greenhouse crop growth environment,is one of the key technologies of modern greenhouse.Controlled object precise mathematical models are the primary form of traditional control methods,a reasonable accurate model of the environment is an important prerequisite greenhouse greenhouse environmental control.Because of the complexity of the greenhouse environment,the modeling process often contains uncertain parameters,therefore,the need for model parameter identification.On the basis of the greenhouse microclimate mechanism model built on the use of PSO algorithm on greenhouse microclimate mechanistic models of temperature and humidity model parameter identification,and to achieve parallelism improved PSO algorithm and based on improved Levy flight,experiments show that the proposed algorithm effectiveness.The main work is based on the above are as follows:In matlab/simulink environment greenhouse mechanism model temperature and humidity model to model.Before-built temperature and humidity model parameter identification,to verify the identification of the model parameters,the model parameters,particle swarm fitness function is set,the sample design,the identification results of error assessment methods,experimental results of performance evaluation the way the relevant design,good preparatory work for the model parameters identification PSO algorithm.On the basis of matlab parallel programming on the realization of the improved PSO algorithm parallelization model parameter identification.Through the parallel design to address the need to recompile calling simulink model and time-consuming big problem,and facilitate effective implementation of the expanded PSO algorithm in cloud computing platforms.Through experiments,compared to stand-alone computing performance in single-core,single multi-core,multi-machine clusters of these three operating environment.The results show that by improving the PSO algorithm parallelization significantly reduces the program running time and improve efficiency.Parallel PSO algorithm is improved by Levi flight.In order to further improve the recognition accuracy,in parallel PSO algorithm based on the Levy flight step of improvements to its velocity updating formula.Through experimental comparison analysis,based on improved PSO algorithm Levi flight,accuracy and stability of the error identification error results generally higher than standard PSO identification result.And through a large number of experiments,the use of the algorithm based on improved flight Levi was determined optimal population size for the particles of the present time model.