Apparatus and method for estimating construction cost based on genetic algorithm

摘要

Disclosed are a construction cost estimation apparatus and method based on genetic algorithm. The database unit stores data on a plurality of influence factors affecting construction costs and a plurality of existing case data including required construction costs. The similarity score calculation unit calculates the similarity score of each of the influence factors of the learning case in which the influence factors and the required construction cost are determined and the influence factors of the plurality of existing case data. The weight generator generates a plurality of weight sets corresponding to a preset number of generations by sequentially performing selection, crossing, and mutation of a genetic algorithm on a weight set having set weights corresponding to each influence factor. The weight determination unit determines a weight set for which the error calculated for the learning case is the minimum for each weight set as the final weight corresponding to each influence factor. The construction cost calculation department selects similar cases from the existing case data based on the similarity scores of each of the influence factors of the plurality of existing case data and the influence factors of each of the new cases to calculate the construction cost and the final weight corresponding to each influence factor. The construction cost of the new case is calculated based on the construction cost of the case. According to the present invention, the variation of the estimated construction cost can be minimized by applying a genetic algorithm in calculating the weight for each attribute required for the extraction of similar cases similar to the new case among the existing cases when calculating the construction cost.