改进GSA算法在机器人路径规划中的应用研究

Research on the Application of Improved GSA Algorithm in Robot Path Planning

针对万有引力搜索算法(GSA)收敛精度低和易陷入局部最优等问题,提出一种改进万有引力搜索算法(IGSA)。引入当前迭代次数t动态调整引力常数,以增强算法逃离局部最优能力;基于边界值改进粒子越界处理策略,以保留粒子多样性,提高算法收敛精度;同时,将中垂线算法用于游离粒子的位置更新,以加速游离粒子的收敛;为适应前述策略,提出自适应权重因子更新粒子位置策略,以提高算法的收敛速度。算法在10个基准测试函数上的结果表明,改进算法在稳定性、收敛速度和精度方面具有较大优势。最后将改进算法应用于机器人路径规划,并与其他智能仿生算法进行路径规划仿真对比实验,结果表明本文改进算法规划路径更短、拐点更少、搜索效率更高。

To address the problems that the universal gravitational search algorithm(GSA)is low convergence accuracy and easy to fall into local optimum,this paper proposed an improved gravity search algorithm(IGSA).Firstly,in order to enhance the ability of the algorithm to escape from local optimum,the improved algorithm used the current iteration number t that is introduced to dynamically adjust the gravitational constant.Secondly,in order to preserve the particle diversity and improve the convergence accuracy of the algorithm,the improved algorithm based on based on the boundary value to improve the particle crossing processing strategy.Meanwhile,the improved algorithm used the mid-pipeline algorithm to update the position of the free particles to accelerate the convergence of the free particles.To adapt to the aforementioned strategy,the improved algorithm used the adaptive weight factor to update particle position strategy to improve the convergence speed.And the experimental results of the algorithm on 10 benchmark test functions show that the improved algorithm has greater advantages in terms of stability,convergence speed and accuracy.Finally,the improved algorithm is applied to robot path planning and compared with other intelligent bionic algorithm path planning through simulation experiments.The results show that the improved algorithm has shorter paths,fewer inflection points and higher search efficiency.