A Light Regulation Model Based On The Fusion Of Light Quality And Intensity For Greenhouse Cucumber

Light is the main factor affecting the yield and quality of crop,and adequate light is necessary for the growth of cucumber and the formation of fruit.But the existing light environment regulation models haven’t considered the demand of light intensity and light quality changing with environment,and failed to establish different growth stage model according to the characteristics of different growth stages.What’s more,the accuracy rate of the existing light environment regulation models need to be further improved.In order to solve the these problems,based on the effect on photosynthetic rate from the light intensity,temperature and light quality,using generalized regression neural network,genetic algorithm,quantum genetic algorithm and nonlinear regression method,a light environment regulation model of coupling light quality and intensity is established.The main work and conclusions of this article are as follows:(1)The proposed photo synthetic rate modelingA photosynthetic rate model based on genetic algorithm-generalized regression neural network model is proposed.Since the spread parameter influences the performance for generalized regression neural network,genetic algorithm is used to find the optimal spread parameter for generalized regression neural network.As a result,genetic algorithmgeneralized regression neural network model is established.In seedling stage of the cucumber,mean absolute error,maximum absolute error,mean relative error,root mean square error between predicted photo synthetic rate and tested photosynthetic rate of the optimal value are 0.294μmol/(m2·s),0.918μmol/(m2·s),4.695%,0.355μmol/(m2·s)respectively,and the correlation coefficient is 0.9989.It shows that the accuracy of the genetic algorithmgeneralized regression neural network photosynthetic rate model is better than generalized regression neural network photosynthetic rate model.What’s more,In flowering stage of the cucumber,mean absolute error,maximum absolute error,mean relative error,root mean square error between predicted photosynthetic rate and tested photosynthetic rate of the optimal value are 0.344μmol/(m2·s),1.235μmol/(m2·s),4.694%,0.448μmol/(m2·s)respectively,and the correlation coefficient is 0.9988.It shows that the accuracy of genetic algorithm-generalized regression neural network model photosynthetic rate model is better than generalized regression neural network photosynthetic rate model.(2)Optimization based on quantum genetic algorithmOptimization of photosynthetic rate based on quantum genetic algorithm is carried out.At a certain temperature,quantum genetic algorithm is used to find the optimal photosynthetic rate,corresponding to the optimal light quality and light intensity.Then according to the relation between the command red light,the command blue light and the optimal light quality,the optimal light intensity,we get the optimal light quality and light intensity.As the initial value of the quantum genetic algorithm is randomly assigned,the results of each optimization are different.In order to obtain the accurate optimization results,when coefficient of optimal light quality variation,coefficient of optimal light intensity variation and coefficient of optimal photosynthetic rate variation all are less than 15%,the optimization results are retained.Quantum genetic algorithm optimization and genetic algorithm optimization result and optimal photosynthetic rate are compared.The results show that the quantum genetic algorithm can obtain more accurate photo synthetic rate.(3)The method of light environment regulation modelingLight environment regulation model is built using Smoothing Splines.Experiments are designed to verify the practicability of light environment regulation model.In the cucumber seedling stage,the maximum absolute error,mean absolute error,root mean square error between the theory of optimal value of photosynthetic rate and photosynthetic rate of the optimal value are 0.124μmol/(m2·s),0.043μmol/(m2·s),0.063μmol/(m2·s);In the cucumber flowering stage,the maximum absolute error,mean absolute error,root mean square error between the theory of optimal value of photosynthetic rate and photosynthetic rate of the optimal value are 1.580μmol/(m2·s),0.377 μmol/(m2·s),0.714μmol/(m2·s).As a result,light environment regulation model have high precision and good practicability.