人类理性程度与人机协作绩效:基于计算实验的演化分析

doi:10.12005/orms.2024.0350

摘要/Abstract

摘要： 随着人工智能的兴起,人机协作显著地提高了任务的效率和质量。人类作为人机协作的主要参与者,在协作中起着关键作用,但当前少有研究关注人类主体特质对人机协作绩效的影响。因此,本文基于有限理性理论,整合虚拟博弈模型以及强化学习算法,构建了考虑人类理性程度的人机协作计算实验模型。通过一个移动协作任务的仿真实验,分析了人类理性程度对协作绩效的影响以及不同机器人适应性的情景下人类理性程度对协作绩效的影响。结果表明,人类理性程度能促进人机协作客观绩效与主观绩效的提升,但这种影响作用对于客观绩效而言边际递减,且机器人适应性越高,人类理性程度对协作绩效的影响越明显。该研究将有限理性理论扩展到了人机协作绩效提升领域的研究中,研究结论对人机团队的管理具有一定的实践启示。

关键词: 人机协作, 有限理性, 协作绩效, 计算实验

Abstract: With the rise of artificial intelligence, human-robot collaboration has significantly improved the efficiency and quality of tasks. As the main participants of human-robot collaboration, humans play a key role in collaboration. The current researches on the influencing factors of human-robot collaboration performance are mainly carried out from the two aspects: (1)The influence of robot characteristics on the collaboration performance, including the characteristics, adaptability and degree of automation of robots. (2)The influence of human-robot relationship on collaboration performance, including human-robot trust and shared mind between human-robot, etc. The above researches have considered the influence of robot behavior characteristics on human-robot collaboration performance. However, few studies have explored the influence of human subject traits on human-robot collaboration performance with human as the center. In addition, the researches on human-robot relationship mainly focus on the trust relationship between human and robot, but those on human-robot collaboration and interaction under different robot characteristics are particularly lacking. Therefore, the scientific questions studied in this paper are: What is the influence mechanism of human rationality degree on human-robot collaboration performance? How does this influence mechanism differ in different robot adaptations?
Due to the dynamic and adaptive nature of human-robot collaboration, the existing methods based on physical experiments and questionnaires take cost and time, and are difficult to understand and reveal the evolutionary characteristics of collaboration from a dynamic perspective. Therefore, based on the bounded rationality theory, this paper integrates the virtual game model and reinforcement learning algorithm to establish an agent-based human-robot collaborative computational model considering human rationality, and then, to observe the influence of human rationality degree on objective and subjective performance of collaboration. Objective performance refers to the completion of objective cooperative tasks between human and robot, while subjective performance refers to the subjective evaluation of collaboration by human collaborators. In this paper, Python language is used to program the model and conduct simulation experiments. To avoid the influence of randomness, experiments are used to run 100 times in each experiment with 200 cycles in each time, and the average value of 100 experiments is used for indicator measurement.
Through a simulation experiment of a move collaborative task, the influence of human rationality on collaborative performance and the moderator role of robot adaptability are analyzed. The experimental results show that increasing the degree of human rationality can improve the performance of human-robot collaboration, but this effect marginally diminishes for objective performance. Besides, the higher the adaptability of robots, the more obvious the impact of human rationality on collaboration performance. The research extends the bounded rationality theory to the field of human-robot collaboration performance improvement. And the research conclusions have practical enlightenment for the management of human-robot teams. Improving the cognitive ability of human collaborators, adopting robots with high adaptability and recruiting human groups with a high degree of rationality will lead to dramatic improvements in the overall performance of human-robot collaboration.
The simulation scenario in this paper is based on mobile cooperative tasks. With the diversification of human-robot collaboration scenarios, we will try to further explore other types of human-robot cooperative tasks in the future. In addition, agent-based computational experiment is a simplification of the real world on the basis of ensuring the rationality of the model. Moreover, it cannot truly represent the complex decision-making process of human beings, such as the influence of psychology, physiology, environment, etc., on decision-making. Behavioral experiments involving human beings can be considered to further demonstrate the conclusions obtained in this paper.

Key words: human-robot collaboration, bounded rationality, collaboration performance, computational experiment

中图分类号:

C931

裘江南, 张方芳, 尤玥, 李梦洁, 高双燕. 人类理性程度与人机协作绩效:基于计算实验的演化分析[J]. 运筹与管理, 2024, 33(11): 37-43.

QIU Jiangnan, ZHANG Fangfang, YOU Yue, LI Mengjie, GAO Shuangyan. Human Rationality and Human-Robot Collaboration Performance: Evolutionary Analysis Based on Computational Experiments[J]. Operations Research and Management Science, 2024, 33(11): 37-43.

参考文献

[1] MCALLISTER M P. Reverse engineering social media: Software, culture, and political economy in new media capitalism[J]. New Media & Society, 2016, 18(7): 1190-1195.
[2] ROBERT L P, YOU S. Human-robot interaction in groups: Theory, method, and design for robots in groups[C] //GROUP’14: 2014 ACM Conference on Supporting Groupwork, November 9-12, 2014,Sanibel Island, Florida, USA, New York: Association for Computing Machinery, 2014: 310-312.
[3] 李忆,喻靓茹,邱东.人与人工智能协作模式综述[J].情报杂志,2020,39(10):137-143.
[4] 张含阳.人机协作:下一代机器人的必然属性[J].机器人产业,2016(3):37-45.
[5] SIEGEL M, BREAZEAL C, NORTON M I. Persuasive Robotics: The influence of robot gender on human behavior[C]//2009 IEEE/RSJ International Conference on Intelligent Robots and Systems, October 11-15, 2009, St. Louis, USA. IEEE, 2009: 2563-2568.
[6] NIKOLAIDIS S, HSU D, SRINIVASA S. Human-robot mutual adaptation in collaborative tasks: Models and experiments[J]. The International Journal of Robotics Research, 2017, 36(5-7): 618-634.
[7] HOFF K A, BASHIR M. Trust in automation: Integrating empirical evidence on factors that influence trust[J]. Human Factors, 2015, 57(3): 407-434.
[8] CHEN M, NIKOLAIDIS S, SOH H, et al. Trust-aware decision making for human-robot collaboration: Model learning and planning[J]. ACM Transactions on Human-Robot Interaction, 2020, 9(2): 1-23.
[9] GAO X, GONG R, ZHAO Y, et al. Joint mind modeling for explanation generation in complex human-robot collaborative tasks[C]//2020 29th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN), August 31-September 4, 2020, Naples, Italy. IEEE, 2020: 1119-1126.
[10] YOU S, ROBERT L. Teaming up with robots: An IMOI (Inputs-Mediators-Outputs-Inputs) framework of human-robot teamwork[J]. International Journal of Robotic Engineering, 2017, 2: 003.
[11] LIU B, FU W P, WANG W,et al. Research on cobot action decision-making method based on intuitionistic fuzzy set and game theory[J]. IEEE Access, 2022, 10: 103349-103363.
[12] ZHENG N N, LIU Z Y, REN P J, et al. Hybrid-augmented intelligence: Collaboration and cognition[J]. Frontiers of Information Technology & Electronic Engineering, 2017, 18(2): 153-179.
[13] LIU C, HAMRICK J B, FISAC J F, et al. Goal inference improves objective and perceived performance in human-robot collaboration[C]//THANGARAJAH J, TUYLS K, JONKER C, et al. Proceedings of the 15th International Conference on Autonomous Agents & Multiagent Systems(AAMAS2016). International Foundation for Autonomous Agents and Multiagent Systems, 2016: 940-948.
[14] ASSOCIATION A E, SOCIETY R E, SIMON H A. Theories of Decision-Making in Economics and Behavioural Science[M]. Springer, 1966: 1-28.
[15] 章平,戴燕.个体决策与学习行为:有限理性建模综述[J].南开经济研究,2006(3):116-128.
[16] 张娟,李亚峰,白园博,等.不确定性分析—敏感性分析方法[J].知识经济,2008(8):101-101.
[17] 周志华.机器学习[M].北京:清华大学出版社,2016.
[18] 袁艺,茅宁.从经济理性到有限理性:经济学研究理性假设的演变[J].经济学家,2007(2):21-26.
[19] DELPRETO J, RUS D. Sharing the load: Human-robot team lifting using muscle activity[C]//2019 International Conference on Robotics and Automation (ICRA), May 20-24, 2019, Montreal, Canada. IEEE, 2019: 7906-7912.
[20] 张可喜.日本开发出可帮人搬桌子的机器人[J].科学咨询,2002(7):46-46.