A comprehensive review on multi robot task assignment and path planning: operational constraints, challenges, and potentials of deep reinforcement learning

Al Mahmud, Suaib and Mohd Ibrahim, Azhar and Khan, Mazbahur Rahman (2026) A comprehensive review on multi robot task assignment and path planning: operational constraints, challenges, and potentials of deep reinforcement learning. Applied Soft Computing, 188 (NA). pp. 1-36. ISSN 1568-4946 E-ISSN 1872-9681

	PDF (Review) - Published Version Restricted to Repository staff only Download (6MB) | Request a copy
Preview	PDF (Scopus) - Supplemental Material Download (188kB) | Preview

Abstract

Task allocation and path planning are two highly focused areas in the arena of Multi Robot Systems (MRS) due to their versatile applications. Whether they are treated separately or in conjunction, the operational constraints play a critical role on their performance. This paper presents a comprehensive review on task assignment and path planning of MRS by dedicating a specific focus on operational constraints, and explores the potentials of Deep Reinforcement Learning (DRL) to address them. A detailed insight into formulation and scheme characterization of Multi Robot Task Allocation (MRTA) problems is presented with an extensive review on the solution techniques of task allocation and path planning, among which the learning-based approaches (particularly DRL) were found to be providing notable advantages over traditional ones. The operational constraints are systematically categorized and challenges appearing alongside them are identified. Furthermore, the paper pinpoints existing research gaps in this regard, and ventures into discovering ways to bridge them using state-of-the-art DRL training algorithms, techniques, frameworks, and concepts, which also expose significant potential and research avenues for MRS. Additionally, the computational resources and accelerators for DRL are discussed to appreciate the real-world deployment of DRL-based MRS. Thus, the paper establishes a benchmark for advancing the swift autonomy of MRS.

Item Type:	Article (Review)
Additional Information:	8484/126798
Uncontrolled Keywords:	Multi Robot Systems, Task Assignment, Path Planning, Constraints, Challenges, Deep Reinforcement Learning
Subjects:	T Technology > T Technology (General)
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Engineering > Department of Mechatronics Engineering Kulliyyah of Engineering
Depositing User:	Dr Azhar Mohd Ibrahim
Date Deposited:	14 Jan 2026 11:25
Last Modified:	14 Jan 2026 16:29
Queue Number:	2026-01-Q1600
URI:	http://irep.iium.edu.my/id/eprint/126798

Actions (login required)

View Item