【作者】 郭嘉；

【导师】 郑俊杰；

【作者基本信息】 华中科技大学 ， 岩土工程， 2008， 博士

【摘要】 从计算的角度来看,生物免疫系统是一个高度并行、分布、自适应和自组织的系统,具有很强的学习、识别、记忆和特征提取能力。人工免疫系统是对生物免疫系统的模拟,具有强大的信息处理能力,是继人工神经网络、进化计算之后新的智能计算研究领域,是生命科学和计算机科学相互交叉而形成的学科,目前已成为研究热点。本文介绍了生物免疫系统、克隆选择原理以及人工免疫系统及其典型应用,对人工免疫算法及其在岩土工程中的应用等方面进行了深入研究,主要包括下以内容:(1)介绍了生物免疫系统的一些基础概念、功能和原理;分析了遗传算法早熟收敛的成因、人工免疫算法的基本理论以及常用免疫算法的基本结构和流程。用5个标准测试函数对克隆选择算法、遗传算法的收敛性能进行了测试,结果表明克隆选择算法在不同类型的函数上具有较快的收敛速度和较高的收敛精度。(2)介绍了可靠度理论的基础知识,基于免疫算法原理和可靠指标的几何涵义,提出了计算岩土工程可靠指标和设计验算点的全局优化算法。对于功能函数的非线性和复杂性,避免了繁复的求导数工作。通过两个工程实例,证明了提出的方法进行岩土工程可靠度分析的有效性。(3)结合武汉地区混凝土管桩试桩资料进行了统计学分析和可靠度分析。采用了JC法和免疫算法两种方法进行可靠度指标的计算。在求解过程中,对于功能函数的复杂性,免疫算法仅需要目标函数的函数值。概括了多元复合地基承载力计算方法,并推导出多元复合地基可靠度研究中的极限承载力状态方程。利用MATLAB编制了相应的计算程序。通过算例,采用Monte-Carlo法和免疫算法进行对比计算,结果表明免疫算法在进行多元复合地基承载力的可靠度分析时,收敛速度快于Monte-Carlo法并且结果精度高。(4)单桩极限承载力受众多因素的影响,Q—S曲线表现出复杂的非线性特征。结合武汉地区预应力混凝土管桩静载荷试验资料,运用MATLAB编写了免疫算法程序对预应力混凝土管桩的极限承载力进行了预测分析。计算结果与试验资料基本一致。(5)沉降预测是岩土工程研究领域一个重要课题,也是工程设计阶段值得关心的问题。应用免疫算法对沉降进行预测,并与最小二乘法、Asaoka法和遗传算法的预测结果进行比较。结果表明,免疫算法是进行沉降预测的有效方法。更多还原

【Abstract】 From the information-processing perspective, the immune system is a highly parallel and distributed intelligent system that has learning, recognition, memory, and associative retrieval capabilities. Artificial Immune System, which has powerful information processing ability, simulates Biology Immune System. It is a novel intelligent computing research field following the invention of Artificial Neural Network and Evolutionary Computation, and it is an emergent interdisciplinary research field generated by life science and computer science and has become a hot point.The basic principles of the biologic immune system, the Clonal Selection Priciple, the researches and typical applications of AIS are introduced in this dissertation. The immune algorithm and its applications in geotechnical engineering are studied. The main works of this dissertation are summarized as follows.(1) Some basic concepts, functions and principles of the biological immune system are introduced. Then the causes of the premature convergence of the genetic algorithm (GA), the basic theory of IA, structure and process of the IA are analyzed. To verify the excellence of the immune algorithm, five numerical objective functions are adopted to compare the performance of genetic algorithm and immune algorithm. The simulation results demonstrate that IA has faster convergence speed and higher convergence accuracy.(2) The basic concepts of reliability theory are introduced. Based on the immune algorithm and geometric implication of the reliability index, a global optimization method is put forward to calculate the reliability index. By this method, reliability of the nonlinear and complex performance function of geotechnical engineering can be obtained without derivation. Two practical examples show that the method presented in this dissertation is reliable and accurate in the reliability analysis of geotechnical engineering.(3) Based on the results of static load tests of bored piles in Wuhan area, the statistics and reliability analyses of the bearing capacity of single PC pile are performed by utilizing the limit state equation of dimensionless random variables. The index of reliability of the single PC pile is calculated with both IA method and JC method. By immune algorithm, only the objective function value is used, reliability of the nonlinear and complex performance function of geotechnical engineering can be obtained without derivation. In recently years, multi-column composite foundation is widely used. The classification and the computation methods of bearing capacity of multi-column composite ground are introduced, and the ultimate bearing capacity equation is deduced. The program of computing reliability index is developed in MATLAB. Three practical examples using Monte-Carlo method and immune algorithm show that immune algorithm has the advantages of rapid convergence and high precision in the reliability analysis of bearing capacity of multi-column composite ground.(4) The Q-S curve of single pile has complicated nonlinear features because the ultimate bearing capacity of single pile is affected by many factors. Based on the results of static load test of prestressed concrete pipe piles in Wuhan area, immune algorithm program is developed by MATLAB to predict the ultimate bearing capacity of prestressed concrete pipe piles. Calculated results are in agreement with those from measurement.(5) To predict the settlement of foundations is an important issue in geotechnical engineering and one of the main concerns in design. IA is used in an attempt to obtain accurate settlement prediction. The predicted settlements obtained by IA are compared with those predicted by the least squares fitting method, Asaoka method and GA. The results show that IA is a useful technique for predicting the settlement of foundations.更多还原