基于多种群协同进化的多逃逸者围捕任务分配-《计算机技术与发展》

文章信息/Info

Title:: Task Allocation of Multi-escapee Roundup Based on Multi-population Coevolution

作者:: 高子璇¹ ; 张国富¹; 2; 3; 4 ; 苏兆品¹; 2; 3; 4 ; 李磊¹; 1. 合肥工业大学计算机与信息学院,安徽合肥 230601;
2. 大数据知识工程教育部重点实验室(合肥工业大学),安徽合肥 230601;
3. 智能互联系统安徽省实验室(合肥工业大学),安徽合肥 230009;
4. 工业安全应急技术安徽省重点实验室(合肥工业大学),安徽合肥 230601

Author(s):: GAO Zi-xuan1 ; ZHANG Guo-fu1; 2; 3; 4 ; SU Zhao-pin1; 2; 3; 4 ; LI Lei1; 1. School of Computer Science and Information Engineering,Hefei University of Technology,Hefei 230601,China;
2. Key Laboratory of Knowledge Engineering with Big Data ( Hefei University of Technology) ,Ministry of Education,Hefei 230601,China;
3. Intelligent Interconnected Systems Laboratory of Anhui Province ( Hefei University of Technology) ,Hefei 230009,China;
4. Anhui Province Key Laboratory of Industry Safety and Emergency Technology ( Hefei University of Technology) ,Hefei 230601,China

Keywords:: swarm robot escape roundup; multi - escapee task allocation; genetic algorithm; multi - population coevolution; staticobstacle avoidance

摘要:: 群机器人逃逸围捕一直是人工智能和机器人领域的研究热点之一。在面向多逃逸者时,如何为每个逃逸者高效地分配合适的机器人以完成协同围捕是一个难点问题。已有研究大都采用距离优先分配的策略,为每个逃逸者选择离它最近的一组机器人进行围捕,在逃逸者数量较多的情况下,难以实现围捕任务的均衡分配,降低了系统围捕的效率。为此,提出了一种基于多种群协同进化的多逃逸者围捕任务分配算法。首先,构建了一种全方向的群机器人逃逸围捕任务分配数学模型;然后,基于遗传算法和多种群协同进化提出了一种多逃逸者围捕任务分配算法,设计了相应的编码方式、交叉和变异策略;最后,在开发的群机器人逃逸围捕仿真平台上测试了算法的有效性。对比实验结果表明,所提算法在完成围捕任务所耗费的步数上最多降低了 20% ,围捕效率最大提高了 25% 。

Abstract:: Swarm robot escape roundup has been one of the research hotspots in the field of artificial intelligence and robotics. Whenfacing multiple escapees,it is a difficult problem to efficiently assign the appropriate robots to each escapee to complete collaborativeroundup. Most of the researches have adopted the distance-first allocation strategy to select the nearest group of robots for each escapee,which makes it difficult to achieve a balanced distribution of the fencing task when the number of escapee is large and reduces theefficiency of system fencing. To this end, a multi - escapee roundup task allocation algorithm based on the co - evolution of multipleswarms is proposed. Firstly,an all-directional swarm robot escape roundup task assignment mathematical model is constructed,and thena multi-escapee roundup task assignment algorithm is proposed based on genetic algorithm and multiple swarm co-evolution,and the corresponding coding methods,crossover and variation strategies are designed. Finally,the effectiveness of the proposed algorithm is testedon the developed swarm robot escape roundup simulation platform. Comparative experimental results show that the proposed algorithmreduces the number of steps consumed in completing the roundup task by up to 20% and improves the roundup efficiency by up to 25% .