基于多目标遗传算法的传感器优化布点研究

针对目前桥梁模态试验中传感器布点优化都是基于单一准则下的研究的局限性,提出采用双准则或多准则下的传感器优化布点研究,该方法改进了以往布点优化中单一目标最优的限制,转而寻求多准则下的优化布点的满意解,而非真正严格数学意义上的最优解。采用Pareto遗传算法,设计了相应的遗传算子和编码方案,成功求解了双准则下传感器最优布点问题,优化结束时,成功给出了一组pareto最优解的前沿面,进而证明了提出的方法是可行的。

OPTIMAL SENSOR PLACEMENT BASED ON MULTI-OBJECT GENETIC ALGORITHM

The present optimal sensor placement is based on single-object criterion in the bridge health monitoring and structure modal test. A new method for optimal sensor placement is presented based on multi-criterion. Using the multi-object optimization algorithm, the satisfying solution (not necessarily optimal in the sense of strict mathematics) of the sensor placement is carried out using Pareto genetic algorithm. The genetic operator and coding precept are designed in this paper. Front side of Pareto solution is obtained at the end of the optimal calculation, proving the feasibility of the proposed method.