Multi-Robot Path Planning Based on an Improved Bag Spider Algorithm
TL;DR
An improved bag spider algorithm for coordinated robot path planning with an adaptive exploration–exploitation schedule that adjusts based on iteration progress and population diversity and an obstacle-aware repair operator that locally bypasses infeasible segments and smooths the resulting paths is introduced.
Abstract
Multi-robot path planning is a complex, multi-constraint optimization problem where multiple robots must reach their targets while avoiding obstacles and each other. Metaheuristic planners are appealing for these challenges because they can explore large continuous spaces without relying on an explicit cost landscape model. However, they often face issues like premature convergence and unstable collision handling in cluttered environments. In this paper, the authors introduced an improved bag spider algorithm for coordinated robot path planning. Two complementary mechanisms are proposed: (a) an adaptive exploration–exploitation schedule that adjusts based on iteration progress and population diversity and (b) an obstacle-aware repair operator that locally bypasses infeasible segments and smooths the resulting paths. Three variants are tested—adaptive, enhanced obstacle avoidance, and a combined version—and compared with the original bag spider algorithm and particle swarm optimization in a 2D environment with five robots and multiple circular obstacles.