Use of genetic algorithms for optimal policies of MR in a bridge network

摘要

In available Bridge Management System (BMS) there are different models to optimize the funds that each Agency dedicates to those matters. The paper presents a model developed for the BMS of the Chiapas State in Mexico, using a joint optimization of the maintenance and rehabilitation policies. The specific transition probabilities of the Markov matrices are estimated for the particular conditions of the Chiapas bridge stock condition state and traffic loading and environmental loads in the area. The optimization problem is solved via a computer application using genetic algorithms (GA) to find the minimum costs for the different maintenance and rehabilitation policies generated by the model. It is shown how the application of the proposed model leads to better budget allocation and less total cost when compared with the standard method of maintenance used till now by the Chiapas State Agency.