A combination of the Euclidean and the Manhattan distance, designed for use in path finding algorithms as the heuristic distance function.
The idea is that Euclidean distance has better results in mostly empty areas and Manhattan in more filled areas (rectangle shapes), so a combination can be created based on the coverage of the area in each scenario.
The a parameter describes the ratio of the accessible area to the total area or simply the coverage of the area.
akrdis(x1, y1, x2, y2, a) = sqrt(pow2(x2-x1) + pow2(y2-y1))*(1-a) + (abs(x2-x1) + abs(y2-y1))*a
akrdis(x1, y1, x2, y2, a) = euclidean(x1, y1, x2, y2)*(1-a) + manhattan(x1, y1, x2, y2)*a