Multi-agent system optimisation in factories of the future: cyber collaborative warehouse study

• Published in: International journal of production research Vol. 60; no. 20; pp. 6072 - 6086
• Main Authors: Dusadeerungsikul, Puwadol Oak, He, Xiang, Sreeram, Maitreya, Nof, Shimon Y.
• Format: Journal Article
• Language: English
• Published: London Taylor & Francis 18.10.2022; Taylor & Francis LLC
• Subjects: Automation 5.0; Collaboration; Collaborative Control Theory (CCT); Completion time; cyber augmentation; cyber collaborative protocol; Factories; Industrial robots; Internet of Things; Multiagent systems; Operating costs; Optimization; Supply chains; Task Administration Protocols (TAP); Warehouses; Workflow
• ISSN: 0020-7543; 1366-588X
• DOI: 10.1080/00207543.2021.1979680

The rapid advancement of technologies leading to automation 5.0 has challenged manufacturers preparing for factories of the future, including warehouses, which are considered a key element in supply chains. Because of technologies such as warehouse robots, Internet of Things, Internet of Services, and cyber-augmented collaboration, the traditional warehouse system structure has been changed, improving its performances significantly. The challenges, however, are how to design a system with multi-agents and technologies to reach maximum potential. In this study, a new collaborative workflow protocol for cyber collaborative warehouse, called Collaboration Requirement Planning protocol for HUB-CI (CRP-H), is developed for optimising the collaborative workflow of a warehouse multi-agent system. The two phases of CRP-H are designed to answer questions: (1) Which robot(s) should execute which task? and (2) When should this task be executed? Results show (with statistical significance) that under CRP-H, total operational cost reduces by 11.84%, and total weighted completion time reduces by 37.11%. When the system has unplanned requests, CRP-H can still reduce total operational cost by 5.70% and total weighted completion time by 10.11%. Lastly, CRP-H, which enables a human input integrated into the design, can also reduce the total operational cost even when critical information is missing.