Optimum repairing of bridge painting systems based upon neural network and genetic algorithms

The aim of this paper is to develop an optimum decision supporting system fbr proposing rational maintenance plan, especially repainting plan, of a group of many existing steel plate- and box-girder bridges. The optimum plan that repaints some steer bridges among many existing ones is determined, selected: of a group of many deteriorating bridges by solving the combinatorial optimization problems with discrete variables and a discontinuous objective function within annual budget. A simple Genetic Algorithms(GA), which is used in a method for searching for most suitable repainting program of steel bridges is adopted to the combinatorial optimization problems. Some numerical examples show to prove the applicability-and usefulness of this decision supporting system for searching for the fittest repainting plan of damaged existing bridges.