The Genetic Algorithm-based Optimal Solution of Water Balance

Abstract:
During the industrial production, it is necessary to calculate the water balance, that is, Sigmainflow = Sigmaoutflow. The paper adopts neural network to train the inflow and outflow of a node and parameters of other production conditions, and then employs genetic algorithm to encode and initialize the inflow of the node and parameters of other production conditions. Afterwards, the present study inputs the initial population into the neural network to run a projection, makes a selection according to the fitness function, and proceeds to a reselection based on the crossover variation of the selected individual, in order to calculate the optimal water balance in this node.