利用遗传及免疫算法进行桥梁维修管理计划的优化

OPTIMIZATION OF MAINTENANCE PLANNING FOR EXISTING BRIDGES USING GA AND IA

以桥梁的耐久性和耐荷性为桥梁的质量指标,建立了考虑桥梁经济性和质量指标的维修加固模型。探讨了简单遗传算法(SGA)和免疫算法(IA)求最优桥梁维修管理计划的解的可能性,并对比了二者解的差别。结果表明当维修计划期间较短时,二者均可给出适当的优化解,当桥梁维修管理计划期间较长时,SGA法得出的准优化解频度减少,解的可靠性降低;而IA法可以给出SGA法无法得出的长期间维修计划的优化解,并给出复数的维修管理计划方案供用户选择。

An optimization system for bridge maintenance planning is developed and included in the bridge intelligent decision support system. In this study, an optimization model for bridge maintenance planning considering both maintenance cost minimization and quality maximization is established by adopting load-carrying capability and durability as the bridge rating indices. The optimal solutions of the maintenance planning obtained using the Simple Genetic Algorithm （SGA） and Immune Algorithm （IA） are compared. It is found that both could provide appropriate optimal solutions, as the maintenance period is shorter. However, the reliability of the solution using SGA decreases gradually as the maintenance period becomes longer. On the other hand, not only can the IA provide more accurate solutions than the SGA, but also can supply plural semi-optimal solutions for user＇s selection.