The Study, Based On Case-based Reasoning Of High Loess Slope Stability Analysis Method

With implication of the strategy of the development of West China, the construction of the highways (especially express highways) has become a priority, which could drive the economic of the West China. On Jan 13th, 2005, the State Department issued The State Plan of the Networks of the Express Highways, from which we can see that a new round of the express highway has just begun in the middle and western area of China, especially the western areas. A serious problem in the express highway construction in the loess areas is that great number of the high loess slope are collapsed after or in the process of construction which would subsequently threaten the lives and safety of the people. Based on the analysis of the traditional approaches, and begins from the intelligent approach, which are Rough Set Theory, Case-based Reasoning, Genetic Algorithm, the paper tries to find a reasonable appraisal approach for the stability evaluation of the high loess slopes. The following results have been achieved through the research:1. Through live surveys of the high loess slopes in the highways being built or already built in western loess areas and the analysis of the basic forms and the mechanism of the distortion and destruction of the high loess slopes, the author concludes that the basic forms of the distortion and destruction are: slope destruction and dome deformation, of which the latter accounts mostly in the loess areas, which is expressed in the form of dome erosion, flaking off. Flaking off happens in the slope where there is soil exposed, while the dome erosion mainly happens in the area where the slopes are from by the new loess. Factors, internal factors, external factors and the human factors, which influence the deformation of the loess have also been analysed.2. As there exists lots of subjective uncertainties, on the discussion of the Rough Set Theory in the stability analysis of the high loess slopes, the paper apply the 18 high loess slopes to simplifying and predicting the result, which showed that the Rough Set Though has its special function in simplifying the properties values and the excess property of each samples. The predicted results fit the reality.3. As the appraisal information of the high loess slope has the character of incomplete and uncertain, based on the case-based reasoning approach in the artificial intelligence, for the...