Article: Cultivating collaborative innovation ability model in higher education based on multi-agent system Journal: International Journal of Computational Systems Engineering (IJCSYSE) 2026 Vol.10 No.9 pp.16 - 26 Abstract: This research utilised a multi-agent system model in higher education to represent students, educators, administrators, and external resource providers as intelligent agents, each exhibiting distinct qualities and behavioural norms to enhance collaborative innovation. A state evolution mechanism was incorporated into a task-oriented collaborative learning process to facilitate the dynamic alteration of agent knowledge acquisition, communication, and problem-solving capabilities. The multi-faceted assessment encompassed task completion rates, inter-role communication, and systemic problem-solving abilities. Simulations and empirical assessments throughout five collaboration phases indicated an average task completion rate of 92.4% and a systematic problem-solving ability score of 90.6, with maximum student-teacher interactions reaching 18 and feedback quality rated at 86. During periods of high demand, the model reduced ability disparities among diverse learners, enhanced adaptability, and preserved operational efficiency. The findings indicate that the paradigm fosters dynamic collaboration and the cultivation of strategic skills, offering scalable and adaptive innovation training alternatives for higher education. Inderscience Publishers - linking academia, business and industry through research

Title: Cultivating collaborative innovation ability model in higher education based on multi-agent system

Authors: Junli Yang

Addresses: Department of Earth and Science Engineering, Shanxi Institute of Technology, Yangquan, Shanxi, 045000, China

Abstract: This research utilised a multi-agent system model in higher education to represent students, educators, administrators, and external resource providers as intelligent agents, each exhibiting distinct qualities and behavioural norms to enhance collaborative innovation. A state evolution mechanism was incorporated into a task-oriented collaborative learning process to facilitate the dynamic alteration of agent knowledge acquisition, communication, and problem-solving capabilities. The multi-faceted assessment encompassed task completion rates, inter-role communication, and systemic problem-solving abilities. Simulations and empirical assessments throughout five collaboration phases indicated an average task completion rate of 92.4% and a systematic problem-solving ability score of 90.6, with maximum student-teacher interactions reaching 18 and feedback quality rated at 86. During periods of high demand, the model reduced ability disparities among diverse learners, enhanced adaptability, and preserved operational efficiency. The findings indicate that the paradigm fosters dynamic collaboration and the cultivation of strategic skills, offering scalable and adaptive innovation training alternatives for higher education.

Keywords: higher education; collaborative innovation capability; multi-agent system; MAS; collaborative learning process; state evolution mechanism; SEM.

DOI: 10.1504/IJCSYSE.2026.154060

International Journal of Computational Systems Engineering, 2026 Vol.10 No.9, pp.16 - 26

Received: 12 Aug 2025
Accepted: 24 Feb 2026
Published online: 10 Jun 2026 *

Title: Cultivating collaborative innovation ability model in higher education based on multi-agent system

Keep up-to-date