Abstract: Due to comprehensive effects of many factors on the swelling-shrinkage of expansive soils, it is difficult to determine the classification of the swelling-shrinkage grade in the process of expansive soil treatment. In this paper, five key factors were chosen as the assessment indexes to construct a cloud model and carry out the uncertain reasoning on the basis of the classification standards, and the qualitative comment of each assessment index was translated into quantitative value to achieve the transition from a qualitative concept to a quantitative data. Furthermore, according to the optimization theory, the objective weights were obtained by using twenty standard samples produced through stochastic interpolation in each grade interval. In association with the subjective weights, the combined weights were calculated. Finally, the new model was applied to practical projects in consideration of the combined weights and the uncertain reasoning, and the total scores and the swelling-shrinkage grades were acquired. Results show that the classification results fit well with the actual situation. The feasibility and applicability of the proposed model are verified. It offers reference to the classification of swelling-shrinkage grade of expansive soils in similar projects.

Key words: expansive soil, swelling-shrinkage grade, classification, cloud model, uncertain reasoning