Multi-Agent Task Coordination Through Dynamic Peer Collaboration

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing agent modules can only process user requests within their specific domain of expertise, limiting the ability to handle tasks requiring multiple domain-specific capabilities, and existing multi-agent collaboration methodologies are rigid and pre-designed, hindering complete automation of complex tasks.

Innovation Solution

A system and method enabling peer-to-peer natural language-based communication between agents, allowing for automatic selection of optimal agents to collaborate and process user requests that cannot be handled by a single agent, using a computing system to determine recipient agents, generate sub-task plans, and integrate external processing results to produce a final result.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple expert agents are implemented to process specialized tasks, then task processing capability is improved, but system complexity increases

Engineering Contradiction:
Improvetask processing capabilityVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

A coordinator agent is introduced as an intermediary component that manages communication and task distribution among multiple expert agents. The coordinator receives user requests, determines which expert agents are needed, and coordinates their collaboration to produce final results, thereby managing system complexity while maintaining versatility

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system is segmented into specialized expert agents, each handling specific task domains, and a coordinator agent that manages overall task distribution. This segmentation allows complex tasks to be broken down and handled by appropriate specialists without requiring a single agent to master all domains

Inventive Principle:
Principle #1Segmentation

2Ease of manufacture

If pre-designed collaboration methodology is used, then implementation is simplified, but flexibility to handle diverse tasks is reduced

Engineering Contradiction:
Improveimplementation simplicityVSAvoidflexibility
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The collaboration methodology transitions from static pre-designed workflows to dynamic adaptive coordination. The coordinator agent dynamically determines task distribution and collaboration patterns based on the specific characteristics of each user request, allowing the system to adapt to diverse tasks while maintaining implementation simplicity through automated decision-making

Inventive Principle:
Principle #15Dynamics

3Device complexity

If one agent handles all tasks, then system complexity is reduced, but task processing quality deteriorates

Engineering Contradiction:
Improvesystem complexityVSAvoidtask processing quality
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

Different parts of the system (expert agents) are assigned specialized qualities and expertise for specific task domains. Each expert agent possesses localized knowledge and capabilities optimized for its domain, ensuring high processing quality for specialized tasks while the coordinator manages overall system complexity

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4600874B1Method and system for processing task through collaboration between agents
Publication Date: 2026.04.08 SAMSUNG SDS CO LTD
  • EP4600874B1 patent drawingFigure 1
  • EP4600874B1 patent drawingFigure 2
  • EP4600874B1 patent drawingFigure 3

AI summary

The present disclosure relates to a method and system for processing a task through collaboration between agents. According to an aspect of the inventive concept, there is provided a method for processing a task through collaboration between agents, which is performed by a computing system. The method may comprise receiving a user request by a first agent from a user terminal and determining whether the first agent may process the user request by itself, inquiring detailed information on each of a plurality of agents when the first agent determines that it cannot process the user request by itself, and determining a recipient of a first task processing request, which requests to process at least some of a plurality of subtasks constituting a target task for generating a processing result for the user request, based on the detailed information, transmitting the first task processing request to a second agent determined as the recipient of the first task processing request, receiving first external processing result data from the second agent in response to the first task processing request and generating final processing result data for the user request by using the first external processing result data and transmitting the final processing result data to the user terminal.