.png){kind=icon}
返回
Strength Learning Particle Swarm Optimization for Multiobjective Multirobot Task Scheduling
.png){kind=icon}
PRE
AI
DOI：10.1109/TSMC.2023.3239953.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Cooperative heterogeneous multirobot systems have attracted increasing attention in recent years. They use multiple heterogeneous robots to execute complex tasks in a coordinated way. The allocation of heterogeneous robots to cooperative tasks is a significant and challenging optimization problem. However, little work has gone into scheduling large-scale cooperative tasks with precedence constraints and multiple conflicting optimization objectives. Existing methods are insufficient to address the issue. We propose a multiobjective model and develop strength learning particle swarm optimization (SLPSO) to optimize multiple objectives. In this article, the problem is converted into a two-step problem of task permutation construction and robot subset selection. In order to coordinate with the time-extended property of the problem, SLPSO utilizes a hybrid encode scheme: an element-based representation for task permutations and a binary representation for robot coalitions. A strength learning strategy with heuristic information guides particles to enhance their best-performing objectives for improving swarm convergence. In addition, an estimation-based local search is developed to improve spare solutions for enhancing swarm diversity, which determines the search direction by estimating fitness improvements. Experimental results on thirty problem instances are elaborated to demonstrate that the proposed SLPSO significantly outperforms the state-of-the-art algorithms in terms of inverted generational distance and hypervolume metrics. The proposed SLPSO can obtain a set of high-quality and diversified solutions.
Keyword:
Task analysis
Robots
Robot kinematics
Multi-robot systems
Search problems
Particle swarm optimization
Metaheuristics
Cooperative
evolutionary computation
multiobjective
multirobot systems
multirobot task scheduling
particle swarm optimization (PSO)
期刊
IF:
10.5
论文数:
1.1W
被引数:
5.0W
机构
引用论文
.png){kind=icon}
分享收藏Multi-objective multi-robot path planning in continuous environment using an enhanced genetic algorithm基于改进遗传算法的连续环境多目标多机器人路径规划
PREAI
分享收藏
分享收藏
查看更多内容