Adaptive Multi-Model–Human Game: A Unified Framework for Self-Adaptive Switching Between Artificial Intelligence Model Games and Human Games

Game theory provides a fundamental framework for understanding strategic interactions, competition, cooperation, negotiation, coordination, and decision-making among rational or boundedly rational participants. With the development of large language models (LLMs), strategic interactions increasingly involve multiple artificial intelligence models capable of playing games, developing strategies, negotiating, forming alliances, evaluating opponents, and adapting their behavior through repeated interaction. Meanwhile, human participants remain essential sources of strategic judgment, behavioral diversity, contextual understanding, social preferences, and real-world decision-making. Existing research has investigated multi-agent games, human–AI games, behavioral game theory, LLM-based strategic reasoning, and interactions between artificial agents and human participants. However, these research directions do not, by themselves, establish a general mechanism-independent architecture in which multi-model games and human games constitute independently available strategic mechanisms that can be adaptively selected, switched, combined, and reconfigured throughout an evolving game process. This paper introduces the Adaptive Multi-Model–Human Game (AMMHG) framework as a specialized theoretical extension of the Universal Mutual-Human Model (UMHM). The proposed framework establishes three fundamental game configurations: Multi-Model Game (M), Human Game (H), and Multi-Model–Human Game (M–H). Multi-Model Game represents strategic interaction among two or more distinct artificial intelligence models, whereas Human Game represents strategic interaction among human individuals or groups. The combined configuration integrates artificial intelligence models and human participants into a shared strategic environment. Unlike conventional architectures that permanently assign strategic interactions to artificial agents or human participants, AMMHG introduces Adaptive Game Mechanism Selection, Adaptive Game Switching, Adaptive Game Combination, and Adaptive Game Reconfiguration. Its central principle is that the same strategic function may be performed through multi-model interaction, human interaction, or their combination, depending on the evolving characteristics of the game. A formal game representation, adaptive switching policy, contextual utility model, transition mechanism, theoretical propositions, architectural design, evaluation methodology, and application scenarios are developed. The framework additionally distinguishes strategic interaction from strategic simulation, model-generated human behavior from actual human participation, and configuration switching from ordinary changes in game strategy. To the best of our knowledge, this paper is the first to formally articulate this particular unified, function-general, and adaptively switchable architecture for multi-model games and human games, specifically integrating independent and combined game configurations with adaptive selection, switching, combination, and structural reconfiguration. The contribution concerns this specific architectural integration rather than the invention of game theory, multi-agent competition, human games, or human–AI interaction.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-19
DOI
https://doi.org/10.5281/zenodo.24362601
Primary Topic
Game Theory and Applications
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Adaptive Multi-Model–Human Game: A Unified Framework for Self-Adaptive Switching Between Artificial Intelligence Model Games and Human Games

Jincheng Zhang
Zenodo (CERN European Organization for Nuclear Research)
Game Theory and Applications
preprint

Adaptive Multi-Model–Human Game: A Unified Framework for Self-Adaptive Switching Between Artificial Intelligence Model Games and Human Games

Jincheng Zhang
preprint en

Abstract

Game theory provides a fundamental framework for understanding strategic interactions, competition, cooperation, negotiation, coordination, and decision-making among rational or boundedly rational participants. With the development of large language models (LLMs), strategic interactions increasingly involve multiple artificial intelligence models capable of playing games, developing strategies, negotiating, forming alliances, evaluating opponents, and adapting their behavior through repeated interaction. Meanwhile, human participants remain essential sources of strategic judgment, behavioral diversity, contextual understanding, social preferences, and real-world decision-making. Existing research has investigated multi-agent games, human–AI games, behavioral game theory, LLM-based strategic reasoning, and interactions between artificial agents and human participants. However, these research directions do not, by themselves, establish a general mechanism-independent architecture in which multi-model games and human games constitute independently available strategic mechanisms that can be adaptively selected, switched, combined, and reconfigured throughout an evolving game process. This paper introduces the Adaptive Multi-Model–Human Game (AMMHG) framework as a specialized theoretical extension of the Universal Mutual-Human Model (UMHM). The proposed framework establishes three fundamental game configurations: Multi-Model Game (M), Human Game (H), and Multi-Model–Human Game (M–H). Multi-Model Game represents strategic interaction among two or more distinct artificial intelligence models, whereas Human Game represents strategic interaction among human individuals or groups. The combined configuration integrates artificial intelligence models and human participants into a shared strategic environment. Unlike conventional architectures that permanently assign strategic interactions to artificial agents or human participants, AMMHG introduces Adaptive Game Mechanism Selection, Adaptive Game Switching, Adaptive Game Combination, and Adaptive Game Reconfiguration. Its central principle is that the same strategic function may be performed through multi-model interaction, human interaction, or their combination, depending on the evolving characteristics of the game. A formal game representation, adaptive switching policy, contextual utility model, transition mechanism, theoretical propositions, architectural design, evaluation methodology, and application scenarios are developed. The framework additionally distinguishes strategic interaction from strategic simulation, model-generated human behavior from actual human participation, and configuration switching from ordinary changes in game strategy. To the best of our knowledge, this paper is the first to formally articulate this particular unified, function-general, and adaptively switchable architecture for multi-model games and human games, specifically integrating independent and combined game configurations with adaptive selection, switching, combination, and structural reconfiguration. The contribution concerns this specific architectural integration rather than the invention of game theory, multi-agent competition, human games, or human–AI interaction.

Zenodo (CERN European Organization for Nuclear Research)
Game Theory and Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon