Selecting a preferred electrical supply system development solution based on neural networks technology

The decision on the development of electrical supply systems is based on the analysis of a large amount of information, comparison of multiple options, and the evaluation of the solution long-term impact. This complicates the selection process for the alternative to the development of the electrical network (ADEN). Thus an automated decision support system, facilitating the determination of the preferred ADEN, is required. The article presents the selection procedure for the mathematical apparatus to solve the specified task, and substantiates its characteristics. It is suggested to use the technology of artificial neural networks (ANN), based on the developed multi-criteria system for estimating ARES, which allows ranking the alternatives based on the degree of their preference. The architecture of the ANN, the algorithm for optimizing weights, and their efficiency at various parameters are substantiated. The F-measure and the percentage of correctly accepted decisions were selected as efficiency indicators. They amounted to 0.9794 and 97.83 %, respectively, for optimal network parameters. The received INS was successfully tested in the software package.