创新科技2025，Vol.25Issue(7)：45-57,13.DOI:10.19345/j.cxkj.1671-0037.2025.7.4
演化博弈视角下人机协同创新的交互冲突与突破路径
Interaction Conflicts and Resolving Pathways of Human-machine Collaborative Innovation:An Evolutionary Game Perspective
摘要
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.关键词
人机交互冲突/人机协同创新/演化博弈/索洛悖论/资源配置理论Key words
human-machine interaction conflict/human-machine collaborative innovation/evolutionary game/Solow's paradox/resource allocation theory分类
管理科学引用本文复制引用
任宗强,陆毅然,苏中原..演化博弈视角下人机协同创新的交互冲突与突破路径[J].创新科技,2025,25(7):45-57,13.基金项目
浙江省软科学研究计划重点项目"工业互联网平台赋能浙江制造业高质量创新的机理、路径与政策保障研究"(2024C25028). （2024C25028）
