Stochastic ant agent for priority-based coverage

In this paper, we present a stochastic ant agent algorithm that provides priority-based coverage of an area. The algorithm gives an agent the capability to concentrate on areas that need cleaning with more dirty areas receiving more rigorous cleaning regimes than less dirty areas. This capability is useful for vacuuming, containing pollution or infection especially in areas of very large dimensions. This results in energy-saving benefits for such an agent due to the priority-based nature of the algorithm. We also show in this work that even though the algorithm is stochastic, it is capable of achieving guaranteed coverage of polluted areas. Using a brief mathematical analysis, we show that the algorithm is capable of tracking spatiotemporal functions as well, provided that the tracking speed is faster than the rate of change of the spatiotemporal quantity. Using the commonly used deterministic Boustrophedon cellular decomposition coverage method as a benchmark, we show that our algorithm can converge to a contaminant´s spatial distribution faster than the time taken to scan the entire environment.