SI-Based mACO Multi-Agent Path Planning in Web3D Mountain Battle Scenes

In mountain battlefield, the path planning of the army needs many detail scene information which based on the semantic information (SI) of the mountain terrain. This paper basing on SI addresses a kind of mACO(mountain Ant Colony Optimal) path planning method to compute an optimal path in the mountain terrain for Web3D. A series of solutions proposed inhere: abstracting the mountain information including ray stone, round wood, entrenchment, bamboo nails trap and the fire power semantic information, then mapping every weapon's destruction value into every mountain terrain and use the value to compute the pheromones, which will be reused in the process of path planning. And at last, a leader-follower idea is used for the army marching basing on the optimal path. The mACO algorithm is compared with the classical ACO algorithm and the A star algorithm, and the experimental results show that the mACO algorithm is real-time and its optimal path is shorter than the others.