Parameter Identification Procedures In Geotechnical Engineering With Computational Intelligences And Their Applications

The solution of inverse problem in geotechnical engineering usually requires optimization of an objective function describing the difference between measured and simulated data. The inverse problem is formulated as optimization problem by defining an objective function. With the improvement of computational intelligences and the enhancement of measuring precisions, the parameter identification in geotechnical engineering has been developed. Based on the identified parameter values, the original design parameters will be modified and improved.Most optimization algorithms used for parameter estimation in geotechnical engineering are gradient-type methods that have the disadvantages of being very sensitive to the initial guesses of parameters and being prone to convergence to local minima. The shortcomings of gradient-based optimization methods lies in that they can not converge into global optimum value of the objective function because there are measuring errors and model errors. Tihonov proved that if the solution of the forward problem is linear in the parameters, then solutions of inverse problem exists, in unique, and depends continuously on the measurement data. The numerical computation results show that the objective function of inverse problem is non-convex while the measurement errors are consider into.Genetic algorithm is a global optimum procedure that is based on Darwin's evolutionary theory. The drawback of genetic algorithm is premature characteristic. The simulated annealing is combined with genetic algorithm. The new tournament selection strategy is proposed. The chosen probability of every individual is equal to each other at starting period. The diversity of population is guaranteed and the premature problem of simple genetic algorithm is overcome. With the population evolutionary, the temperature of simulated annealing algorithm decreases and the chosen probability of having good solution individual will increase.Artificial neural networks have self-adaptive, self-organization and leaning abilities. The modified BP algorithm is presented by optimization of leaning step-size. The convergence rate of neural network is improved and oscillation problem is effectively solved during the iteration process. The numerical simulation results show, compared with traditional BP algorithm, the convergence rate and identification precision can be improved.Compared with gradient-based optimization methods, simulated annealing algorithm is recognized to have better capability to find the global optimal solution. The inverse problem of identifying aquifer parameters is treated as a combinational optimization problem. The numerically computational results show that the procedure developed in the paper is capable of dealing with both unknown heat transfer coefficient and unknown surface temperature, and has ability to fitting measurement noise. The ill-posedness of the inverse problem as characterized by instability and non-uniqueness is overcome by using simulated annealing algorithm. The effectiveness and flexibility of presented inversion technique are evaluated and compared with descent search method.Ant colony optimization is applied into the parameter identification of geotechnical engineering and its application domain is developed. The ant colony optimization is presentedto identify the transmissivity and storage coefficient for a two-dimensional, unsteady state groundwater flow model. The convergence rate of ant colony optimization is slow because the algorithm finds the solutions by using pattern search. To speed up the rate of convergence and enhance inversion precision, the simulated annealing is applied to help ant colony optimization. The new simulated annealing-ant colony optimization is presented. The new hybrid inversion algorithm has the advantages of information evaporation of ant colony optimization and the speed searching characteristic of simulated annealing.Based upon the dividing-work characteristic of natural ants, the scout ant is applied to ant colony optimization. The scout ant finds better...