The paper presents a semantic model for optimizing the design and interaction of intelligent services within the critical information infrastructure of a smart city. In the context of the increasing complexity and interdependence of critical information infrastructure systems, the proposed model integrates formal engineering methods with semantic mechanisms of knowledge coordination. The work formalizes many key parameters of optimality and introduces quantitative indices for architecture assessment. To manage design uncertainty, an entropy approach is used that formalizes the risks of certainty of the result, implementation, and criticality. The semantic model describes the process of dissemination and coordination of meanings between agents (experts, documents, models) using a diffusion equation based on the conceptual framework of cognitive categories. The semantic coherence index quantitatively evaluates the correspondence of descriptions to the subject area. Modeling of complex behavior and interdependencies of components of the critical information infrastructure is carried out using a modified extended open hybrid automaton. The integration of a semantic model, a hybrid automaton and a Markov decision process into a modular platform for supporting management decisions allows for a well-reasoned choice of the optimal architecture.