Abstract: To effectively promote the development of disruptive technological innovation within the context of future manufacturing,a tripartite evolutionary game model including advanced manufacturing enterprises,research institutions and government departments was constructed.Based on the stability analysis of each subject's strategies,the impact of factors such as government incentives and penalties,and R&D costs on system evolution were examined.A set of replicator dynamic equations was established to analyze the stability of the equilibrium points of the game system.Reinforcement learning algorithms were employed to cross-validate the system's robustness.The results indicated that the system development could be divided into four stages:nascent,development,expansion and maturity.The system demonstrated strong robustness under various operational conditions,and the government's decision-making trajectory exhibited significant amplitude variations.The increase in the proportion of innovation revenue distribution for manufacturing enterprises had a positive incentive effect on the system's evolution rate during the expansion and maturity stages.Similarly,the independent R&D costs of research institutions had a positive effect during the nascent and expansion stages.However,enterprise investment in technology outsourcing had a negative constraint throughout the entire process.

Key words: future manufacturing, disruptive technological innovation, evolutionary game, reinforcement learning

摘要： 为更好地促进未来制造背景下制造业颠覆性技术的创新发展,构建包括先进制造企业、研发机构与政府部门的三方演化博弈模型。基于主体策略稳定性分析,探究政府奖惩力度、技术研发成本等因素对系统演化的影响;建立复制动态方程组,分析博弈系统均衡点稳定性;采用强化学习算法交叉检验系统鲁棒性。结果表明,系统发展分为新生期、发展期、扩张期与成熟期4个阶段;系统运行在不同条件下均具有较好的鲁棒性,且政府部门的实际决策过程轨迹振幅较大。制造企业创新收益分配占比的提升在扩张期与成熟期对系统演化速率有正向激励作用,研发机构的独立研发成本在新生期与扩张期呈现类似效果,技术外包投入则在全过程中起反向约束作用。

关键词: 未来制造, 颠覆性技术创新, 演化博弈, 强化学习