Cognition of decision behavior based on belief state: a case study of implementation in CGF simulation system
Abstract In the field of modern military warfare, decision behavior modeling of Agents is a critical approach for simulating the decision-making process of various military entities in battlefield environments. However, existing models suffer from instability, lack of flexibility and adaptability, a...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2025-06-01
|
| Series: | Complex & Intelligent Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s40747-025-01948-z |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract In the field of modern military warfare, decision behavior modeling of Agents is a critical approach for simulating the decision-making process of various military entities in battlefield environments. However, existing models suffer from instability, lack of flexibility and adaptability, and difficult to generate effective decisions when facing complex battlefield environments characterized by high dynamics, incomplete information, and large state-action spaces. Therefore, this paper addresses the decision behavior modeling of Agents in above battlefield environments. Firstly, decision behavior of Agents in dynamic environments is time continuous, the UCTCD algorithm is incorporate into the planning method; Secondly, some information is unobservable in some scenarios, an environmental state sampling method named SBS-UCTCD is proposed, the most likely estimate of the current battlefield state is obtained by maintaining a single belief state; Finally, the large tree search space in complex battlefield environments makes it difficult to make decisions in limited time, a hybrid decision behavior modeling method HS-UCTCD is proposed by combining HTN planning and SBS-UCTCD, decomposing high-level tasks through HTN planning and searching optimal actions in refined tasks through SBS-UCTCD, thereby enhancing decision efficiency and reducing decision time. The performance of the proposed environment state sampling and decision behavior planning methods are validated respectively with different AIs and on different maps on the microRTS platform. Experiments show that the proposed decision behavior model enables rapid decision making and effective execution in response to changes in the environment and is more versatile. |
|---|---|
| ISSN: | 2199-4536 2198-6053 |