An effective initialization method for genetic algorithm-based robot path planning using a directed acyclic graph

• A new initialization method for GA-based robot path planning is proposed.
• A graph creation process for generating multiple feasible paths is proposed.
• The performance comparison of initialization methods is conducted.
• The effect of initialization for GA-based robot path planning is demonstrated.

The goal of robot path planning is to find a feasible path that proceeds from a starting point to a destination point without intersecting any obstacles in the given environment. Recently, genetic algorithm-based robot path planning methods have been widely considered in the intelligent robotics community. Because the initialization process significantly influences the performance of the genetic algorithm, an effective initialization method is required. However, investigation on this subject is still lacking. In this paper, we propose an effective initialization method for genetic algorithm-based robot path planning. Experimental results comparing genetic algorithms with conventional initialization methods and the proposed initialization method showed that the proposed method leads to high quality paths in a significantly shorter execution time.