A human-inspired collective intelligence model for multi-agent based system
The collaborative and competitive nature of multi-agent systems (MAS) is visible through the simple social mode of communication that emerges between human-agent interactions or agent-to-agent interactions. A simple mode of communication involves the fundamental actions carried out by individual age...
Saved in:
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Published: |
Universiti Putra Malaysia Press
2023
|
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The collaborative and competitive nature of multi-agent systems (MAS) is visible through the simple social mode of communication that emerges between human-agent interactions or agent-to-agent interactions. A simple mode of communication involves the fundamental actions carried out by individual agents in achieving their desired goal. The sum of these achievements contribute to the overall group goal. Comparatively, the collective intelligence (CI) of a MAS simply means that these agents should work together to produce better solutions than those made possible when using the traditional approach. In designing MAS with CI properties, formalisation of a higher level deliberation process is essential. A high level deliberation process refers to the judgement comprehension of tasks, reasoning and problem solving and planning. In this paper, we propose our Collective Intelligence Model, CIM, which has the potential to control and coordinate a high-level deliberation process of a MAS. CIM is inspired by the emerging processes of controlled discussion, argumentation and negotiation between two or more intelligent human agents. These processes screen and validate the deliberation process through a cross-fertilisation approach. The emergent property of the cross-fertilised ideas results in an intelligent solution that solves optimisation-related tasks. � 2017 Universiti Putra Malaysia Press. |
|---|