Motivation. One of the tasks of nuclear power plants and other potentially dangerous objects monitoring employing unmanned aerial vehicles (UAV) is flying over specified points of the territory of such objects. Finding the optimum routes often involves different options for the traveling salesman problem solving. However, if there are multiple starting positions, there is a need to solve the traveling salesman problem for each variant of the UAV start (for each variant of the start-end point of the route). The subject matter of the paper is the process of minimizing the flight time of visiting the specified points of the potentially dangerous object territory, taking into account the locations and models of the UAV fleet. The tasks to be solved are: to develop an algorithm for determining the optimal route for flight over of the given points of the potentially dangerous object territory for the fleet, each UAV of which is at its separate starting position; to show the possibility of using the proposed algorithm to minimize the flight time for visiting all of the appointed control posts of the automated radiation situation monitoring system for Zaporizhzhia nuclear power plant. The methods used are: graph theory, mathematical optimization models, methods for solving the traveling salesman problem. The following results were obtained. The faceted classification of the traveling salesman problem for UAV flight routing is offered. The steps of the algorithm for determining the optimal route of flight over of the specified points of the potentially dangerous object territory by the UAV fleet are described. The problem of determining the fastest flight over of 11 control posts of the automated radiation monitoring system for Zaporizhzhia nuclear power plant is solved for two cases: 1) UAV "Leleka-100" are at all starting positions, 2) UAV "Leleka-100" is at the first starting position, various modifications of the model "R-100" are at the rest starting position. Changes in the optimal route when changing UAV models and speeds are shown. Conclusions. The results obtained should be used to justify the composition of the UAV fleet, simulate its application, evaluate its target effectiveness, as well as to create algorithmic support and software for ground control station operators’ work places. Further research should focus on developing models that take into account the possibility of refueling UAVs or recharging their batteries at stationary or moving posts while being on a route.
Influence of parameters of the ant colony algorithm on the traveling salesman problem solution
