Path planning as multi-objective optimization using the NSGA-II algorithm

The most important goal of path planning is to find the safest and fastest path between two points in a given environment. In this study, the path planning of a moving robot considers two objective functions: the distance to goals and the number of turning points. This research uses a Multi-Objective Optimization approach by utilizing the results of the A* algorithm to create individuals in the initial population for NSGA-II. The study utilized a 29x29 grid with various kinds of obstacles that must be avoided. The algorithm used is NSGA-II with an initial population using the path generated by the A* algorithm. The genetic operators used are two-point crossover with a probability of 0.9 and polynomial mutation with a probability of 0.01. The simulation results show that the proposed method can produce a path that has fewer turning points and the shortest possible path length. In four study case results, the NSGA-II shows a better performance in reducing the number of turning points. In study case map 1, NSGA-II produces 8 turns while A* produces 13. In study case map 2, NSGA-II produces 8 turns while A* produces 12. Study case map 3 shows significant differences, with NSGA-II producing 26 turns and A* producing 41. Study case 4 yields identical results from both algorithms.