Hybrid genetic algorithm-based approach for estimating flood losses on structures of buildings

Abstract

Occurrence of extreme natural events raises the need for a quick and accurate estimation of losses on the insured residential property in order to support timely recovery of the area. Although various models are now available to estimate the amount of loss on buildings, there is a lack of models providing a sufficient level of detail and accuracy that can be used for insurance purposes. In this study, a hybrid genetic algorithm-based model for flood loss estimation on the structures of buildings is presented. The proposed model combines the ordinary least squares method, a genetic algorithm, and the bill of costs method, which offers a good balance of maximum simplicity on the one hand and the accuracy of calculation on the other hand. The model considers the geometric characteristics (dimensions and shape) of rooms and is enabled to work with various types of materials and structures, as well as a variable depth of flooding. The results achieved show that in quick loss estimation, the model provides highly accurate results which meet the requirements for its use for the purposes of settlement of real insurance claims by insurance companies. The article outlines the potential automated connection of the model to insurance companies' information system in order to create a simple building information model (BIM) of the insured property (building's structures).