Three Dimensional Adaptive Path Planning Simulation Based on Ant Colony Optimization Algorithm

Ant Colony Optimization (ACO) is a metaheuristic widely used to solve different problems. This work proposes a three-dimensional simulation of adaptive path planning. New features were added to the basic ACO algorithm. First, the Random Walk based on visibility for initializing the pheromone matrix. The visibility of a node is the distance from the current node to the target node (d_it) over the distance from the possible node to the target node (d_jt). The second feature is the inclusion of Killer Nodes for adaptive behavior. These nodes remove an ant and execute a decay function that removes the contributions over a wrong path. Finally, several experiments were performed to evaluate the solution accuracy, convergence time, and computational complexity. These results showed that the feasible ACO solution is near to the optimal solution with accuracy over 95% for most cases. It demonstrates that the algorithm provides promising results and finds a route after the addition of dynamic obstacles.