To solve the key problem of simultaneously optimizing path length

smoothness

and comfort in global path planning for guide robots in static environments

and to overcome the limitations of meta-heuristic algorithms in terms of population initialization

convergence efficiency

and adaptability

an improved guided bitterling fish optimization(GBFO)algorithm is proposed.By integrating a heuristic greedy initialization strategy with a smoothness penalty mechanism

the algorithm improves population diversity and convergence efficiency

while optimizing the path length

smoothness

and comfort.Performance comparison experiments on CEC2022 benchmark functions show that the GBFO algorithm exhibits strong convergence and robustness

demonstrating its potential for application in global path planning for guide robots.Comparative experiments in grid map environments show that the GBFO algorithm outperforms both the grey wolf optimization (GWO)algorithm and the original bitterling fish optimization(BFO)algorithm.The GBFO algorithm achieves better performance than the original BFO algorithm in terms of the shortest path length

the number of turns

and computation time.