School of Computer Science, Liaocheng University, Liaocheng 52059, China.
Math Biosci Eng. 2023 Jan;20(2):2501-2529. doi: 10.3934/mbe.2023117. Epub 2022 Nov 23.
Effective path planning (PP) is the basis of autonomous navigation for mobile robots. Since the PP is an NP-hard problem, intelligent optimization algorithms have become a popular option to solve this problem. As a classic evolutionary algorithm, the artificial bee colony (ABC) algorithm has been applied to solve numerous realistic optimization problems. In this study, we propose an improved artificial bee colony algorithm (IMO-ABC) to deal with the multi-objective PP problem for a mobile robot. Path length and path safety were optimized as two objectives. Considering the complexity of the multi-objective PP problem, a well-environment model and a path encoding method are designed to make solutions feasible. In addition, a hybrid initialization strategy is applied to generate efficient feasible solutions. Subsequently, path-shortening and path-crossing operators are developed and embedded in the IMO-ABC algorithm. Meanwhile, a variable neighborhood local search strategy and a global search strategy, which could enhance exploitation and exploration, respectively, are proposed. Finally, representative maps including a real environment map are employed for simulation tests. The effectiveness of the proposed strategies is verified through numerous comparisons and statistical analyses. Simulation results show that the proposed IMO-ABC yields better solutions with respect to hypervolume and set coverage metrics for the later decision-maker.
有效的路径规划（PP）是移动机器人自主导航的基础。由于 PP 是一个 NP 难问题，智能优化算法已成为解决该问题的热门选择。人工蜂群（ABC）算法作为一种经典的进化算法，已被应用于解决众多实际优化问题。在这项研究中，我们提出了一种改进的人工蜂群算法（IMO-ABC）来解决移动机器人的多目标 PP 问题。路径长度和路径安全性被优化为两个目标。考虑到多目标 PP 问题的复杂性，设计了良好的环境模型和路径编码方法，使解决方案可行。此外，应用了混合初始化策略来生成有效的可行解。随后，在 IMO-ABC 算法中开发并嵌入了路径缩短和路径交叉算子。同时，提出了一种可变邻域局部搜索策略和一种全局搜索策略，分别增强了算法的开发和探索能力。最后，使用包括真实环境地图在内的代表性地图进行了仿真测试。通过大量的比较和统计分析验证了所提出策略的有效性。仿真结果表明，所提出的 IMO-ABC 算法在超体积和集覆盖指标方面为决策者提供了更好的解决方案。
