An intelligent approach for optimal control of fruit-storage process using neural networks and genetic algorithms

Optimal control techniques based on fruit responses offer the possibility for the qualitative improvement of fruit during the storage process. This study presents a new intelligent control technique, including neural networks and genetic algorithms, for realizing the optimal control of the fruit-storage process. The control input is a relative humidity h, and the control outputs are two types of fruit responses: the water loss W-h(t) and the development of lesion by fungi D-h(t) of the fruit (h: relative humidity, t: sampling time). An objective function is given by the reciprocal number of the sum of the average values in both W-h(t) and D-h(t). For control, the storage process was divided into 1 steps. First, responses of W-h(t) and D-h(t), as affected by relative humidity, were identified using neural networks. The I-step setpoints of relative humidity which maximize the objective function were then searched for through simulation of the identified model using genetic algorithms. Control results suggested that the storage process should be treated as a dynamic process, and an intelligent approach proposed here is useful for the optimization of such a control process. (C) 1997 Elsevier Science B.V.