Application of neural network to aesthetic design of bridges

The present study proposes to investigate whether the currently accepted criteria for assessing the bridge aesthetics and harmony with the environment are appropriate or not, and to define the factors relevant for the landscape design. An expert system for the quantitative assessment of landscape design of girder and arch bridges was developed. This method makes use of both objective and subjective data of past bridge designs by applying the learning ability of the neural network. These data include, besides the element color, which is important in the environmental design of bridges, the criteria of the bridge harmony with the environment. The main feature of the neural network is pattern recognition, which allows it to learn from the input-output response patterns of the sample data. This ability provides useful information for landscape design