Digital Twin Action Selection for Mixed Individual and Group Agents
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multi-agent simulation technologies struggle to accurately represent selection tendencies of agents when individuals and groups are mixed, leading to decreased simulation accuracy.
Innovation Solution
The simulation apparatus generates a digital twin of the real world, specifies agents corresponding to groups, and uses an action selection model that considers the number of members in each group to predict action selection, employing a multi-agent simulation to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a discrete selection model is used for action selection in multi-agent simulation, then the simulation can be executed efficiently, but the simulation accuracy decreases when individuals and groups are mixed
Solution Approach 1:
The patent segments the action selection model into two distinct components: a discrete selection model for individual agents and a group selection model for group agents. This segmentation allows each model to be optimized for its specific purpose, maintaining computational efficiency while improving accuracy for group-based simulations. The simulation system can selectively apply the appropriate model type based on the agent composition, thus resolving the contradiction between efficiency and accuracy.
Solution Approach 2:
The patent introduces dynamic adaptability by allowing the simulation system to switch between different selection models based on the mixed composition of individuals and groups. The system dynamically determines which model to apply for each agent or group of agents, enabling flexible and accurate simulation execution that adapts to varying simulation conditions while maintaining overall efficiency.
2Device complexity
If a simple discrete selection model is used, then the model complexity is low, but it cannot correctly express selection tendencies when individuals and groups are mixed
Solution Approach 1:
The patent divides the selection modeling into separate discrete selection models for individuals and group selection models for groups. This segmentation maintains relative simplicity within each model type while capturing the specific selection tendencies appropriate to each agent type, avoiding the need for a single overly complex model that would be difficult to manage and interpret.
Solution Approach 2:
The patent changes the parameters and structure of the selection model based on agent type. By adjusting the model parameters to reflect individual versus group characteristics, the system achieves accurate representation of selection tendencies without requiring a completely new complex model framework for each case.
Data Source
Figure 1
Figure 2
Figure 3(a)~3(b)
AI summary
A simulation program causes a computer (10) to execute a process including generating (S101) a digital twin in which a real world is reproduced in a virtual space, specifying (S102. S103), in the generated digital twin, an agent corresponding to a person existing in the real world, and performing (S106) a simulation of an action of the person based on composition information of the specified agent and an action selection model for predicting action selection of the person.