Multi-Agent Movement Control for Balanced Delay and Energy Use

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing multi-agent navigation technologies often result in extreme delays for specific agents due to variations in route lengths and times, as they minimize overall route lengths and times by prioritizing one agent over others, leading to inefficient navigation in environments with multiple agents.

Innovation Solution

An agent control device and method that uses a movement determination unit, a change amount determination unit, and an operation control unit to reduce variations in compensation, delay, and energy consumption by applying outputs from a first model and a second model, which are trained on observation sets of control target agents and surrounding agents, to optimize movement directions, times, and energies.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the sum of route lengths of all agents is minimized, then the overall navigation efficiency is improved, but specific agents may experience extreme delays

Engineering Contradiction:
Improveoverall navigation efficiencyVSAvoiddelay of specific agents
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent changes the optimization parameter from minimizing the sum of route lengths to minimizing the variation (standard deviation) of route lengths among agents. This parameter transformation resolves the contradiction by shifting focus from aggregate efficiency to equitable distribution of travel times, preventing extreme delays for individual agents while maintaining overall system productivity.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If the route length of a specific agent is minimized by increasing its priority, then that agent's navigation efficiency is improved, but the route lengths of other agents increase

Engineering Contradiction:
Improvenavigation efficiency of prioritized agentVSAvoidroute length of other agents
Core Design Contradiction:
ProductivityVSLength of moving object

Solution Approach 1:

The patent applies equipotentiality by making all agents equal in terms of optimization criteria - minimizing the standard deviation of route lengths treats all agents symmetrically. No single agent is prioritized, ensuring that route length adjustments benefit the overall distribution equity rather than favoring specific agents, thus resolving the trade-off between individual and collective navigation performance.

Inventive Principle:
Principle #12Equipotentiality

3Stability of the object's composition

If variations in route lengths among agents are reduced, then equity in navigation is improved, but overall navigation efficiency may decrease

Engineering Contradiction:
Improveequity in navigationVSAvoidoverall navigation efficiency
Core Design Contradiction:
Stability of the object's compositionVSProductivity

Solution Approach 1:

The patent uses feedback mechanisms through reinforcement learning where agents receive rewards or penalties based on how their route lengths contribute to the overall standard deviation. This feedback loop allows the system to dynamically adjust routes to minimize variation while maintaining reasonable overall efficiency, resolving the contradiction by providing continuous optimization guidance that balances equity and productivity.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP4343476A1Agent control device, learning device, agent control method, learning method, and computer-readable recording medium
Publication Date: 2024.03.27 OMRON CORP
  • EP4343476A1 patent drawingFigure 1
  • EP4343476A1 patent drawingFigure 2
  • EP4343476A1 patent drawingFigure 3

AI summary

An agent control device includes a movement determination unit configured to input a first observation set, obtained by observing a state of a control target agent, and to input a second observation set, obtained by observing a state of at least one other agent in a periphery of the control target agent, to a first model and to determine information regarding movement of the control target agent in accordance with an output of the first model, a change amount determination unit configured to input the first observation set and the second observation set to a second model and to determine a change amount with respect to the information regarding movement of the control target agent, and an operation control unit configured to operate the control target agent by applying the change amount determined by the change amount determination unit to the information regarding movement determined by the movement determination unit.