Comparison of Genetic Algorithm with Numerical Methods for Solving the Problem of Radio Navigation Reference Point Placement in a Local Navigation System

This article presents a comparison of numerical methods for solving the problem of optimizing the placement of radio navigation reference points in a local navigation system. This paper compares the performance of 3 numerical methods for the problem of automated placement of radio navigation reference points in a local navigation system: genetic algorithm, particle swarm method and Monte Carlo method. The article contains descriptions of the algorithms operation, comparison of methods by the algorithm operation time, by the final number of radio navigation reference points, as it directly determines the cost of a local navigation system organization, and by the quality indicator - by the average value of radial mean-square error over the whole room for different positioning methods (ToF, AoA, TDoA, RTDoA, AoA - ToF, AoA - TDoA, AoA - RTDoA). As a result of the experiments, it can be stated that the genetic algorithm copes with the task of finding a quasi-optimal arrangement of radio navigation reference points at a given room-averaged radial mean-square error faster than other algorithms presented in the paper, while involving a smaller number of radio navigation reference points.