Search

Multi-robot path planning (MRPP) in continuous and known environment is studied in this paper via proposing a novel local path planning approach. To plan optimal collision-free paths for multiple robots simultaneously, a novel implementation method suitable for the meta-heuristic algorithms is devised, and an improved artificial bee colony (ABC) algorithm is developed. Three enhancements to the ABC algorithm are made in this context. Firstly, to better lead the search direction, the global best individual is involved in the search equations of employed bee phase and scout bee phase. Meanwhile, to boost exploitation capability, the learning method of teaching-learning based optimization (TLBO) algorithm is incorporated into the onlooker bee phase. The proposed learning-based ABC (ABCL) algorithm is used to determine the subsequent positions for all the robots based on their current coordinates considering the path length, safety and planning efficiency. The experimental studies on benchmark functions show that ABCL is outstanding in solving different types of optimization problems compared against seven effective meta-heuristic algorithms. More importantly, MRPP simulation results prove that ABCL outperforms its competitors in terms of generating optimal collision-free paths and running time. Compared with the original ABC, ABCL improves these two aspects on average for all tasks by 2.1% and 12.6%, respectively. Therefore, with the proposed implementation method, ABCL can be considered as an effective MRPP solution. • A metaheuristic-based approach is proposed for the multi-robot path planning problem. • A learning-based ABC algorithm is developed to solve time-sensitive online problems. • Simulations of different MRPP tasks were completed, validating the designed approach. [ABSTRACT FROM AUTHOR]

Artificial bee colony (ABC) algorithm is a popular optimization algorithm with excellent exploration ability and various applications. Nevertheless, its effectiveness is limited by the one-dimensional search strategy. Therefore, in order to improve its performance, a reinforcement learning (RL) based ABC algorithm is proposed (named ABC_RL). In ABC_RL, the number of dimensions to be updated in search equation of the employed bee phase is varied and adjusted intelligently via RL. Moreover, two improved search strategies are adopted to maintain a nice balance between diversification and intensification. The performance of ABC_RL is evaluated through a series of comparisons conducted on CEC 2017 benchmark problems. The results indicate that ABC_RL outperforms the compared ABC variants considering the solution accuracy. In addition, a robot path planning problem is concerned to further test the effectiveness of ABC_RL. And the comparison results show the advantages of ABC_RL in terms of path length and running time. • A reinforcement learning based artificial bee colony algorithm is proposed. • The frequency of perturbation is adjusted intelligently. • Two enhanced solution search equations are developed. • A type of heavy-tailed distribution is used to improve randomness. • Experiments and application show nice performance of the proposed ABC_RL algorithm. [ABSTRACT FROM AUTHOR]

Artificial bee colony (ABC) is a prominent algorithm that offers great exploration capabilities among various meta-heuristic algorithms. However, its monotonous and one-dimensional search strategy limits its searching performance in the solving process. Thus, to address this issue, a Q-learning based multi-strategy integrated ABC algorithm (QMABC) is proposed. In the QMABC, multiple search strategies are proposed to utilize different individual experiences and search approaches for solution updates. Then, Q-learning is employed for strategy selection. In comparison to previous studies, this paper introduces more effective state and action configurations within the framework of Q-learning. To evaluate the performance of the QMABC, CEC 2017 benchmark functions are adopted to compare it to different meta-heuristic algorithms including ABC based and non-ABC based algorithms. Moreover, applications in path planning are implemented to further verify the effectiveness of the QMABC. Overall, it should be highlighted that the proposed QMABC demonstrates superiority in both numerical and practical experiments. • A Q-learning based multi-strategy integrated ABC algorithm is proposed. • Multiple strategies are introduced to the onlooker bee phase. • Q-learning is utilized to adaptively determine the most suitable strategy. • Novel state and action settings are designed in Q-learning framework. • Proposed algorithm shows superiority in numerical and practical experiments. [ABSTRACT FROM AUTHOR]