Dynamic Multi-Robot Task Allocation under Uncertainty and Temporal Constraints

被引:0
作者
Choudhury, Shushman [1 ]
Gupta, Jayesh K. [1 ]
Kochendeefer, Mykel J. [1 ]
Sadigh, Dorsa [1 ]
Bohg, Jeannette [1 ]
机构
[1] Stanford Univ, Stanford, CA 94305 USA
来源
ROBOTICS: SCIENCE AND SYSTEMS XVI | 2020年
基金
美国国家科学基金会;
关键词
ASSIGNMENT; ROBUST; BREAKDOWNS; ALGORITHMS; TAXONOMY; MACHINE; SEARCH; JOBS;
D O I
暂无
中图分类号
TP24 [机器人技术];
学科分类号
080202 ; 1405 ;
摘要
We consider the problem of dynamically allocating tasks to multiple agents under time window constraints and task completion uncertainty. Our objective is to minimize the number of unsuccessful tasks at the end of the operation horizon. We present a multi-robot allocation algorithm that decouples the key computational challenges of sequential decision-making under uncertainty and multi-agent coordination and addresses them in a hierarchical manner. The lower layer computes policies for individual agents using dynamic programming with tree search, and the upper layer resolves conflicts in individual plans to obtain a valid multi-agent allocation. Our algorithm, Stochastic Conflict-Based Allocation (SCoBA), is optimal in expectation and complete under some reasonable assumptions. In practice, SCoBA is computationally efficient enough to interleave planning and execution online. On the metric of successful task completion, SCoBA consistently outperforms a number of baseline methods and shows strong competitive performance against an oracle with complete lookahead. It also scales well with the number of tasks and agents. We validate our results over a wide range of simulations on two distinct domains: multi-arm conveyor belt. pick-and-place and multi-drone delivery dispatch in a city.
引用
收藏
页数:10
相关论文
共 56 条
[1]   Dynamic capacity acquisition and assignment under uncertainty [J].
Ahmed, S ;
Garcia, R .
ANNALS OF OPERATIONS RESEARCH, 2003, 124 (1-4) :267-283
[2]   Robust and stable flexible job shop scheduling with random machine breakdowns using a hybrid genetic algorithm [J].
Al-Hinai, Nasr ;
ElMekkawy, T. Y. .
INTERNATIONAL JOURNAL OF PRODUCTION ECONOMICS, 2011, 132 (02) :279-291
[3]   Better bounds for online scheduling [J].
Albers, S .
SIAM JOURNAL ON COMPUTING, 1999, 29 (02) :459-473
[4]   On-demand high-capacity ride-sharing via dynamic trip-vehicle assignment [J].
Alonso-Mora, Javier ;
Samaranayake, Samitha ;
Wallar, Alex ;
Frazzoli, Emilio ;
Rus, Daniela .
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA, 2017, 114 (03) :462-467
[5]  
[Anonymous], 2015, DECISION MAKING UNCE
[6]   Suboptimal Variants of the Conflict-Based Search Algorithm for the Multi-Agent Pathfinding Problem [J].
Barer, Max ;
Sharon, Guni ;
Stern, Roni ;
Felner, Ariel .
21ST EUROPEAN CONFERENCE ON ARTIFICIAL INTELLIGENCE (ECAI 2014), 2014, 263 :961-+
[7]  
Behrens JK, 2019, IEEE INT CONF ROBOT, P8705, DOI [10.1109/icra.2019.8794022, 10.1109/ICRA.2019.8794022]
[8]  
Bertsekas D., 2005, Dynamic Programming and Optimal Control
[9]   Julia: A Fresh Approach to Numerical Computing [J].
Bezanson, Jeff ;
Edelman, Alan ;
Karpinski, Stefan ;
Shah, Viral B. .
SIAM REVIEW, 2017, 59 (01) :65-98
[10]  
Boutilier C, 1996, THEORETICAL ASPECTS OF RATIONALITY AND KNOWLEDGE, P195