Genetic Algorithm Based Search in Slope Stability Analysis

摘要

Genetic algorithms (GA) -a relatively new search and optimization procedure based on the mechanics of natural selection and evolution that mimics the principles of natural genetics and survival of the fittest rule has been applied to slope stability problem for locating the critical slip surface of a soil slope corresponding to minimum factor of safety. The program starts with a randomly generated population of slip-circle centers- the 'artificial chromosomes' (which consists of a concatenation of binary sub-strings), where every character present is an 'artificial gene'. The 'fitness' of each chromosome is worked out and random pairs of reproduced above average parent chromosomes are crossed over probabilistically in a mating pool at some randomly chosen site to produce offspring. These child chromosomes are further mutated, again probabilistically at each locus to place the better child sub-strings in a new population set. The generated set replaces the initial one and the loop continues with the aim to thrust each generated chromosome towards the global optima. The stability analysis is done integrating GA with Bishop's simplified method of slices. The results are superimposed on an example cited by Spencer (1967) to highlight its powerful features associated with versatility.