关键词: 路径规划, 蚁群算法, 自适应步长, 双向交互, 节点距离指数, 启发式函数

Abstract: In response to the slow convergence rate, poor flexibility, and the question of the ant colony algorithm getting trapped in local optima, a bidirectional interactive multi-step ant colony algorithm is proposed for research on path planning. Firstly, a dual population approach is adopted, where exploration is conducted through both forward and reverse directions using the ant colony, and an adaptive step-length strategy is implemented to address the issues of local optima trapping and the inflexibility of ants. Secondly, the algorithm is optimized using an adaptive ant population strategy and an improved heuristic function, with the node distance index of the bidirectional ant colony guiding the algorithm’s node transitions to accelerate convergence. Experimental simulation data indicates that the bidirectional interactive multi-step ant colony algorithm not only achieves global fast convergence but also exhibits high stability and shorter computational time. The quality and convergence speed of the obtained solution is superior to those of the improved ant colony algorithm and the multi-step ant colony algorithm based on terminal distance indicators compared in the reference.

Key words: path planning, ant colony algorithm, adaptive step-length, bidirectional interaction, node distance index, heuristic function