基于多约束场景的BFO-ACO漫游路径规划

1.深圳大学计算机与软件学院,广东深圳518060;2.深圳大学信息中心,广东深圳518060

计算机应用;路径规划;多约束;漫游;死锁;蚁群算法;细菌觅食算法;虚拟环境

BFO-ACO roaming path planning based on multi-constraint scenarios
LIN Xiaoling1,WANG Zhiqiang1,2,GUO Yanyan2,and ZHU Zexuan1

1.College of Computer Science and Software Engineering, Shenzhen University, Shenzhen 518060, Guangdong Province, P. R. China;2.Information Center, Shenzhen University, Shenzhen 518060, Guangdong Province, P. R. China

computer application; path planning; multiple constraints; roaming; deadlock; ant colony algorithm;bacterial foraging algorithm; virtual environment

DOI: 10.3724/SP.J.1249.2022.04463

备注

目前基于蚁群算法的路径规划用于多约束条件下寻找最优路径时,容易陷入局部最优解并导致收敛速度慢.为此,在路径长度、有效景点区域数量、路径平滑性和路径障碍距离等约束条件下,构造一种适应度函数模型,以评价漫游路径的质量.提出混合细菌觅食优化思想的改进蚁群优化(bacterialforagingoptimizationandantcolonyoptimization,BFO-ACO)算法,采用禁忌表优化策略解决传统蚁群算法的死锁问题,提高算法初期的路径多样性,通过引入细菌觅食算法的复制和驱散机制,提高收敛速度,跳出局部最优值.实验结果表明,BFO-ACO算法可在多约束环境下以较少的迭代次数获得高质量的漫游路径,为漫游路径设计提供了参考.
In recent years, when the path planning based on ant colony algorithm is used to find the optimal path under multiple constraints, it is easy to fall into the local optimal solution and lead to slow convergence. Under the constraints of path length, number of effective scenic spots, path smoothness and path obstacle distance, this paper constructs a fitness function model to evaluate the quality of roaming path. A hybrid algorithm of bacterial foraging optimization and ant colony optimization (BFO-ACO) is proposed by using the taboo table optimization strategy to solve the deadlock problem in the traditional ant colony algorithm, which improves the path diversity in the early stage of the algorithm. The replication and dispersion mechanism of bacterial foraging algorithm is introduced to improve the convergence speed as well as jump out of the local optimum. Experimental results show that the BFO-ACO algorithm obtains a higher-quality roaming path with the fewer iterations under multi-constrained environments, which provides the foundation for designing roaming path.
·