A decision support model for selecting community development projects based on fuzzy logic

Abstract

The article proposes an approach to selecting a priority public project for financing in the context of a limited budget and lack of unanimity among community residents. Based on the results of an expert survey, a fuzzy decision-making model was built that considers three key variables: the amount that the community is willing to invest, the expected utility of each project, and the decision on the feasibility of increasing or decreasing funding. The Mamdani fuzzy inference system in the MATLAB environment was used to implement the model. Nine basic rules were formed based on linguistic assessments and the results were defuzzied.

To build an analytical model of the system's behavior, the method of full rotatable design of the second-order experiment was used, which allowed obtaining a quadratic regression dependence of the output variable on the logarithm of the project cost. The voting results showed the absence of a clear leader among the projects, which further substantiates the feasibility of using the proposed approach. The decision to select the optimal project was made based on minimizing the absolute value of the variable Y, which reflects the degree of deviation from the financial balance.

The proposed tool is relevant for the processes of decentralization and development of local self-government, as it allows for formalizing decision-making in situations of limited resources and competition between alternatives.