Computer-supported evaluation for seismic performance of existing buildings

A seismic performance evaluation is one of the highly complicated multi-criteria evaluation problems. This study presents a hierarchy of performance criteria that consists of performance categories such as strength, stiffness, configuration, and age of structures at a higher level while their subcategories considering the capacity of subsystems and components at lower levels mainly focus on the stage of preliminary evaluation. Two methods for determination of performance index are proposed; a simple composition method and a fuzzy inference method. Both methods use weighting factors to represent relative importance of multi-criteria in the determination of performance index. The simple composition method that operates on definite index values and assumes independence between performance criteria is easy to use but difficult to address the cases where complicity and uncertainty are of important problems. The fuzzy inference system uses fuzzy concepts where uncertainties are inherently unavoidable due to insufficient information and engineering assumption and exploits fuzzy rules in the multi-criteria evaluation by replicating heuristic knowledge and experience in logic-based descriptive rules. Moreover, the single performance index obtained by defuzzification enables us to draw a representative and comprehensive seismic performance index from a set of individual performance indexes at multi-levels.