CAJ | 학술논문

基于自然进化原理的遗传算法是一种解决岩土工程复杂非线性问题十分有前景的方法。然而遗传算法容易陷入局部收敛，对遗传算法进行改进加入退火因子解决了这一问题。软土地基沉降受众多因素的影响，在时序上表现出复杂的非线性特征。针对木兰溪防洪工程的沉降预测问题，运用MATLAB编写模拟遗传退火算法程序进行了预测分析。计算结果与实测资料基本一致，它在全局优化、非线性优化、多参数优化等方面表现出了传统算法无法比拟的优势。
기우자연진화원리적유전산법시일충해결암토공정복잡비선성문제십분유전경적방법。연이유전산법용역함입국부수렴，대유전산법진행개진가입퇴화인자해결료저일문제。연토지기침강수음다인소적영향，재시서상표현출복잡적비선성특정。침대목란계방홍공정적침강예측문제，운용MATLAB편사모의유전퇴화산법정서진행료예측분석。계산결과여실측자료기본일치，타재전국우화、비선성우화、다삼수우화등방면표현출료전통산법무법비의적우세。
Genetic algorithm based on natural evolution is a prospective solution for complicated nonlinear issues of geotechnical engineering. However this algorithm is easy trapped into local convergence and the improved one by annealing factor'adding can resolve this problem. Soft soil foundation settlement, affected by many factors, presents complicated nonlinear feature at time sequence. In connection with settlement prediction for Mulanxi flood-prevention project, MATLAB is used to write a program simulating genetic annealing algorithm to analyze the predication. The calculating result is consistent with the measured data. Therefore, the genetic algorithm has much more superiority than the traditional one in respect of overall optimization, nonlinear optimization, muhiparameter optimization.