A hybrid co-evolutionary genetic algorithm for multiple nanoparticle assembly task path planning

In this paper, a hybrid co-evolutionary genetic algorithm (HCGA) has been presented for determining the optimal moving paths of several nanoparticles in a complex environment. In the proposed approach, an artificial potential field (APF) has been used to determine the feasible initial paths for moving the nanoparticles. The proposed APF prepares a potential map of the environment using the initial positions of the nanoparticles and positions of the obstacles and surface roughness. The cost function used in this paper includes the area under the critical force-time diagram, surface roughness as well as path smoothness. The performed investigations indicate the importance of each of these parameters in determining the optimal paths for displacing the nanoparticles. Also, the dynamic application of crossover and mutation operators has been used to avoid premature convergence. Using the information of the potential map, two new operators have been introduced to improve the feasible and infeasible paths. Furthermore, a novel co-evolutionary mechanism for solving the multiple nanoparticle path planning problems has been presented. The proposed co-evolutionary mechanism is able to determine the best destination for each particle, optimal sequence of moves for several particles, and also the optimal path for moving each particle. Finally, the performance of the proposed HCGA has been compared with the conventional genetic algorithm (CGA) and the ant colony optimization algorithm.