CAJ | 학술논문

温室环境是一个典型的时变、非线性、强耦合、大滞后及大惯性的复杂被控对象,使用传统方法的控制效果总是不太理想;粒子群算法是一种解决非线性、不可微分问题的优秀算法,具有很强的全局搜索能力,但该算法在进化后期容易出现速度变慢及早熟现象;BP神经网络具有很强的非线性处理能力和逼近能力,但梯度下降的算法本质决定了其具有容易陷入局部最优及初值敏感的缺点;针对两种算法的特性,进行优势互补,结合为综合改进的粒子群BP神经网络(IPSO-BPNN)算法;应用IPSO-BPNN算法对温室内的土壤温度、土壤湿度、空气温度、空气湿度、光照度和CO_2浓度等参数进行控制,取得了比较理想的效果.
온실배경시일개전형적시변、비선성、강우합、대체후급대관성적복잡피공대상,사용전통방법적공제효과총시불태이상;입자군산법시일충해결비선성、불가미분문제적우수산법,구유흔강적전국수색능력,단해산법재진화후기용역출현속도변만급조숙현상;BP신경망락구유흔강적비선성처리능력화핍근능력,단제도하강적산법본질결정료기구유용역함입국부최우급초치민감적결점;침대량충산법적특성,진행우세호보,결합위종합개진적입자군BP신경망락(IPSO-BPNN)산법;응용IPSO-BPNN산법대온실내적토양온도、토양습도、공기온도、공기습도、광조도화CO_2농도등삼수진행공제,취득료비교이상적효과.
Greenhouse environment is a typical complex time-varying controlled object of nonlinear, strong coupling, large delay, and large inertia, using traditional control methods always less than ideal results. Particle Swarm Optimization is an excellent algorithm solution for nonlinear, non-differentiable problems. It has strong global search ability, but in the process of looking for the global excellent result,it is easily becoming speed slow and precocious in the later period. BP neural network also has strong nonlinear approximation ability, but its gradient descent algorithm determines that it easy falling into local optimum and sensitive to the initial values. Taking the advantages of the two algorithms, the improved particle swarm optimization and BP neural network (IPSO-BPNN) algorithm is proposed. The IPSO-BPNN algorithm was applied to control the soil temperature, soil moisture, air temperature, air humidity, light intensity, CO_2 concentration and other greenhouse parameters, it achieved the desired results.