Abstract
When dealing with landslide structural damage assessment, considering that the scale of the analysis is territorial, it is both technically impossible and economically inconvenient to perform generalized and, at the same time, detailed investigations. In these cases, it is therefore necessary to implement various levels of detail, depending on the size of the analysis, the objectives, the available time and the economic constraints. This multilevel approach usually seems to be the most effective and useful because it allows to calibrate the available resources.
The first level of the analysis, necessarily quick and relatively superficial, should be extended to all the buildings of the territory under study. The objective is to identify those buildings that require a deeper investigation. In this scenario, models based on classical structural mechanics seem to be difficult to apply for a rapid crack pattern diagnosis. This is why a new approach, based on a simple model (i.e. the Load Path Method, LPM), has been proposed by Palmisano and Elia (2013). However, according to Palmisano and Elia (2013) the LPM is very useful for rapidly searching the ‘most plausible solution’ instead of the exact solution. To find the solution, optimization algorithms are necessary. In this scenario, this article aims to show how the Bi-directional Evolutionary Structural Optimization method by Yang et al. (1999) can be very useful to optimize the strut-and-tie models obtained by using the Load Path Method.