Using "biological" genetic algorithms to solve the travelling salesman problem with applications in medical image processing

This paper explores the use of biologically motivated inversion operators to solve the travelling salesman problem (TSP). A TSP coding method is suggested and is used to investigate the "evolution" of contours in medical X-ray images. Contour "patterns" have been obtained for up to 32,000 generations and the "evolution" of the contour patterns has been observed for generation steps of 8,000. Initial results have shown that the technique could be a powerful tool for contour extraction, even in the case of an image containing more than a single contour.