---

In the first impounding of rockfill dams, additional settlements occur in upstream side in saturated rockfills due to collapse
phenomenon even high rainy seasons can cause additional deformation in the dumped rockfills. Unfortunately these
displacements are not taken into account in the conventional numerical models which are currently used to predict embankment
dam behavior during impounding. In this paper to estimate these displacements, strain hardening-strain softening model in Flac
is modified based on the laboratory tests, in which same impounding process in such dams is considered. Main feature of the
model is reproduction of nonlinear behavior of rockfill material via mobilized shear strength parameters and using collapse
coefficient to display induced settlement due to inundation. This mobilization of shear strength parameters associated with some
functions for dilatancy behavior of rockfill are used in a finite difference code for both dry and wet condition of material. Collapse
coefficient is defined as a stress dependent function to show stress release in the material owing to saturation. To demonstrate
how the model works, simulation of some large scale triaxial tests of rockfill material in Gotvand embankment dam is presented
and results are compared with those from laboratory tests, which are in good agreement. The technique could be used with any
suitable constitutive law in other coarse-grained material to identify collapse settlements due to saturation

---

The present paper focuses on size optimization of scallop domes subjected to static loading. As this type of space structures includes a large number of the structural elements, optimum design of such structures results in efficient structural configurations. In this paper, an efficient optimization algorithm is proposed by hybridizing particle swarm optimization (PSO) algorithm and cellular automata (CA) computational strategy, denoted as enhanced particle swarm optimization (EPSO) algorithm. In the EPSO, the particles are distributed on a small dimensioned grid and the artificial evolution is evolved by a new velocity updating equation. In the new equation, the difference between the design variable vector of each site and an average vector of its neighboring sites is added to the basic velocity updating equation. This new term decreases the probability of premature convergence and therefore increases the chance of finding the global optimum or near global optima. The optimization task is achieved by taking into account linear and nonlinear responses of the structure. In the optimization process considering nonlinear behaviour, the geometrical and material nonlinearity effects are included. The numerical results demonstrate that the optimization process considering nonlinear behaviour results in more efficient structures compared with the optimization process considering linear behaviour. .