Identification of hydraulic parameters for unsaturated soils using particle swarm optimization

Determination of material parameters for unsaturated soils from laboratory or field tests can be difficult due to the large number of parameters required for many constitutive models. With increasing computing power readily available, parameter search using modem optimisation procedures is now feasible. In this study the identification of hydraulic parameters from the back analysis of a transient infiltration problem is illustrated. Particle Swarm Optimization (PSO) is utilized in the search for the optimal set of parameter values. Two approaches are described: one where a limited set of the parameters is sought and the second where the whole set is sought. For the latter it is shown that a multi-step range contracting method is appropriate and leads to a computationally economic solution.