With the accelerated replacement of traditional fuel vehicles by new energy electric vehicles (EVs)

the rational deployment of charging infrastructure has become critical to urban sustainability. However

the dynamic uncertainties in EV operating states

urban evolution trends

and operator competition significantly increase the complexity of planning. To achieve dynamic adaptation between planning strategies and urban development

this paper proposes an adaptive dynamic planning strategy for EV charging infrastructure from a government perspective. First

a physical system simulation model is constructed

incorporating user behavior

operator profit evaluation

and government policy constraints

so as to dynamically assess investment benefits of station deployment. Second

this paper designs a fixed-point iteration-based algorithm to efficiently compute the steady- state user demand and regional electricity price equilibrium. Furthermore

this paper formulates a Markov decision process to optimize the construction process and the layout of charging stations. Finally

this paper proposes a spatiotemporally aware Monte Carlo tree search reinforcement learning algorithm to address the uncertainty challenges posed by urban evolution scenarios. Case studies demonstrate that the proposed method outperforms baseline algorithms in terms of social welfare

achieving a performance improvement of 14.04%.