The artificial neural network (ANN) method is used to study the macroscopic model of an actual water distribution system. For the first time, the Ant Colony Optimization (ACO) algorithm is implemented to optimize the node numbers of the hidden layers in the ANN model. The ANN model contains two hidden layers with a maximum of 64 nodes per layer. Each node number in the hidden layers is transformed into a binary representation using Gray coding. In doing this, the logical structure of the ACO algorithm is altered from one of two decision points with sixty-four paths per point to one of twelve decision points with two options per point. This newly defined logical structure makes better use of the parallel nature of the ACO algorithm. Careful preparations of the input data used in the ANN model are made. The study indicates that the ANN method is an attractive alternative to the conventional regression analysis method in modeling water distribution systems.
[1]
Robert DeMoyer,et al.
Macroscopic Distribution‐System Modeling
,
1975
.
[2]
C. T. Haan,et al.
Calibration assessment and data collection for water distribution networks
,
2001
.
[3]
J. Shi.
REDUCING PREDICTION ERROR BY TRANSFORMING INPUT DATA FOR NEURAL NETWORKS
,
2000
.
[4]
Andrzej Bargiela,et al.
Neural Networks Based Decision Support in Presence of Uncertainties
,
1999
.
[5]
Marco Dorigo,et al.
Ant system: optimization by a colony of cooperating agents
,
1996,
IEEE Trans. Syst. Man Cybern. Part B.