For the long-term safety of the medium and small bridges, a lightweight monitoring method[1] was proposed to monitor several important parameters of bridge structure. However, a lot of factors need to be considered when selecting the bridge for monitoring in a complex road network. Therefore, a suitable multi-objective optimization method needs to be proposed for the decision-making of lightweight monitoring of medium and small bridges.
In this paper, a decision-making method for medium and small bridge lightweight monitoring based on multi-objective optimization by NSGA-ll[2] was established. In this decision-making system, the general goal was to achieve the maximum network performance and minimum network cost, and the constraint conditions were the minimum individual performance and maximum network cost. To decrease the dimensions of decision vectors, the decision variables were decomposed by a random grouping strategy under the cooperative coevolution framework[3]; In addition, tabu search was introduced for accelerating convergence and reducing verbosity, and fuzzy membership degree was proposed to obtain the optimization results.
Finally, this optimization method was used in the lightweight monitoring decisions of a county bridge network. The distribution of the Pareto solution set in the process of evolution was analyzed. It is proved to be continuous and well-distributed, indicating a good astringency of this method. Then the objective function performance of NSGA-ll method and AHP method were calculated. Compared with the AHP method, NSGA-ll method can reduce the overall monitoring cost by 8.0% under the same network performance and constraint conditions, which proves that the optimization result with a lower overall cost can be obtained by NSGA-ll method. The feasibility and effectiveness of the optimization method based on NSGA-II are verified. Therefore, this method can be adopted to make scientific and reasonable decisions for the lightweight monitoring of medium and small bridges. It can also provide a reference for some similar engineering practices concerning multi-objective optimization strategy.