伪并行遗传算法及其在桥梁维修管理优化中的应用  

Pseudo Parallel Genetic Algorithm and Its Application to Optimization of Maintenance Planning for Existing Bridges

作　　者：王砚如[1] 王桂萱[1] 

机构地区：[1]大连大学土木工程技术研究与开发中心,辽宁大连116622

出　　处：《计算机系统应用》2010年第5期154-159,共6页Computer Systems & Applications

基　　金：日本文部科学省科学研究项目(10450169)

摘　　要：遗传算法(Genetic algorithms,简称GA)是模拟生物在自然环境中的遗传和进化过程而形成的一种全局优化概率搜索算法。但该算法有时存在着早熟现象,导致搜索过早收敛,无法得到全局最优解。为此,分析伪并行遗传算法的优缺点,探讨用伪并行遗传算法(pseudo parallel genetic algorithm,简称PPGA)求最优桥梁维修管理计划的解的可能性,并用于桥梁维修管理计划的优化。实验结果验证了该方法的有效性和GA相比,具有较快的收敛速度,为桥梁维修管理提出了新的方法。Genetic algorithm（GA） is an algorithm used to find approximate solutions to difficult problems through the application of the principles of evolutionary biology to computer science. It can do a global search quickly. But premature phenomenon still exists in application. As a result, the searching converges earlier and the solution of the optimization cannot be got. Therefore, this paper analyzes the merits and demerits of the pseudo parallel genetic algorithm（PPGA） and discusses the optimum possibility by using PPGA, and at last, applies it to the optimization of maintenance planning for existing bridges. The experimental result has confirmed the validity of this method. Contrast to SGA, PPGA is quick in convergence. It prgvides a new method for existing bridges＇ maintenance plan.

关 键 词：桥梁管理 维修管理计划 伪并行遗传算法 约束多目标优化 

分 类 号：U445.7[建筑科学—桥梁与隧道工程]