An Improved Ant Colony Optimization Applied in Robot Path Planning Problem

an improved Ant colony optimization algorithm (PM-ACO for short) is proposed to solve the robot path planning problem. In PM-ACO, ants deposit pheromone on the nodes but not on the arcs, resulting in that the trails of pheromone become the form of marks, which consist of a series of pheromone points. After ant colony’s tours, the iteration-best strategy is combined with an r-best nodes rule to update the nodes’ pheromone. The stability of PM-ACO is analyzed and some advancement to the algorithm is proposed to improve the performance. Because the pheromone on several arcs is integrated into the pheromone on one node, a rapid pheromone accumulation occurs easily. It is the major causes to the instability. An r-best nodes rule is presented for regulating the pheromone distribution and an adaptive mechanism is designed to further balance the pheromone arrangement. In addition, to shorten the time wasted in constructing the first complete solution and get a better solution, an azimuth guiding rule and a one step optimization rule are used in local optimization. By establishing a grid model of the robot’s navigation area, PM-ACO is applied in solving the robot path planning. Experimental results show that an optimal solution of the path planning problem can be achieved effectively, and the algorithm is practical.