All publication are peer-review
Peer review will take the from of double-blind review Judge objectively and impartially
There is no conflict of interest for the reviewer
Review articles shall be kept strictly confidential prior to publication

Industrial Technology Progress

Interaction Conflicts and Resolving Pathways of Human-machine  Collaborative Innovation: An Evolutionary Game Perspective 

Ren Zongqiang1, 2 , Lu Yiran1, 2 , Su Zhongyuan3 

(1.Business School, Wenzhou University, Wenzhou 325035, China; 2.Wenzhounese Economy Research Insti⁃ tute, Wenzhou University, Wenzhou 325035, China; 3.China Telecom Corporation, Wenzhou 325802, China)

Abstract: With the rapid evolution of artificial intelligence technology, human-machine col⁃ laborative innovation has become an important innovation paradigm to promote industrial up⁃ grading and technological breakthroughs. However, in reality, human-machine interaction is  generally characterized by "inefficient collaboration" and "technology mismatch", which maps  out the phenomenon of disconnection between technological inputs and outputs revealed by  "Solow's paradox". This paper focuses on the interest game relationship and resource allocation  mechanism in the process of human-machine collaborative innovation and identifies four key  variables affecting this collaborative innovation: basic collaborative benefit, employee learning  cost, AI arithmetic cost, and benefit distribution ratio. To analyze these factors, the paper adopts  the evolutionary game method to construct a bilateral game model consisting of human employ⁃ ees (cooperative/exclusionary) and intelligent machines (active/passive), and portrays the behav⁃ ioral evolution trajectory under the dynamic matching of cost and benefit. Based on this analy⁃ sis, two types of optimization paths are proposed, namely, "cost control mode" and "benefit in⁃ centive mode". Simulation experiments show that the human-machine collaborative innovation  system has significant threshold effects and nonlinear evolution paths: ①under the revenue in⁃ centive mode, only when the basic collaborative revenue breaks through the critical threshold  can the system stimulate the positive feedback effect and realize the continuous growth of the  collaborative innovation revenue; the revenue incentive mode is more suitable for innovation  scenarios with a high degree of uncertainty in the environment, and it promotes the collabora⁃ tive strategies of the two sides by improving the collaborative dividend, and gradually stabilizes  the convergence of the two sides. ②Under the cost control mode, the employee learning cost  and AI arithmetic cost must be strictly controlled in the appropriate interval to avoid degrada⁃ tion of the strategy due to high cost; the cost control mode is more suitable for innovation envi⁃ronments with strong resource constraints, and can effectively improve the evolution efficiency  of the system by compressing the collaborative cost of the human-machine parties. ③In addi⁃ tion, the combination of the two strategy modes can effectively solve the "cold start" dilemma at  the initial stage of collaborative innovation and accelerate the system to converge to the Pareto  optimal state stably. ④Static allocation is prone to imbalance in long-term evolution, and the  dynamic allocation model can increase the complexity of evolution and adapt to the market envi⁃ ronment by adapting to the change of contribution. Further, this paper constructs two dynamic  optimization models, "dynamic cost coordination model" and "dynamic benefit allocation  model", introduces human-machine collaborative willingness to match the degree of willingness  and willingness-driven benefit allocation mechanism, accurately adjusts the cost allocation, and  dynamically adjusts the incentive ratio. By doing so, it improves both the stability and innova⁃ tion efficiency of the human-machine collaborative innovation system. This research not only  provides a quantitative basis for solving the "cold start" problem of human-machine interaction  but also helps the dynamic expansion of resource allocation theory in intelligent manufacturing  scenarios. 

Key words: human-machine interaction conflict; human-machine collaborative innovation; evolutionary game; Solow's paradox; resource allocation theory