Development of a model for assessing the comfort of an apartment building using “artificial intelligence methods”

Abstract

The object of the study is a model of an apartment in new multi-story buildings, which are evaluated by customers in the process of their purchase. It has been practically established that one of the most problematic areas is the processing of the results of subjective evaluations of customers who give their impressions of a particular apartment and wish to determine its preliminary value. To eliminate this problem, the authors proposed to use an approach using a Petri net, which is characterized by the dynamics of changes due to the application of a genetic algorithm. The obtained result, which is expressed both qualitatively and quantitatively, makes it possible to minimize the use of financial and human resources in the process of concluding a contract for the purchase and sale of an apartment. This is due to the fact that the ambiguity during processing of the model proposed by the authors is eliminated using the method of fuzzy logic derivation according to the Mamdani algorithm. This approach has a number of features: the ability to work with subjective data, to generalize them, without involving a complex mathematical apparatus. Thanks to this, an increase in the financial capabilities of the construction company is ensured due to an increase in the transparency of the formation of the price for an apartment in a new building, a reduction of unnecessary costs by almost 90%. On the part of the client, the cost of time and money to attract additional services of specialists is reduced by more than half. This approach will help the client to analyze the comfort factors of living in the apartment he wants to buy. The developed model is proposed to be used in the fields of construction and real estate services. This will make it possible to improve the conditions of buying and selling an apartment in a new building. The condition for the practical use of the obtained results is the commissioning of the building in which the apartment is evaluated. And also a simple and friendly system interface in which the model studied by the authors will be implemented.