The article continues the author's previous research in the field of building a model for operational planning of cargo operations in a seaport. It covers a wide range of issues (stratification, semiotics, artificial intelligence). The main attention is paid to modeling on the upper "mathematical" stratum-volume-calendar planning based on multi-stage nonlinear and dynamic programming (NP), heuristics and extrapolation. The optimal plan should be sent to the lower stratum, where cargoplans, cargo technological schemes are considered, and modeling tools - artificial intelligence systems. The upper stratum models are based on dispatch-demurrage, steel time, and the rate of simultaneous processing. Three groups of models are proposed. First, n-stage separable NP problems; an algorithm for solving a two-stage NP problem by dynamic programming with reduced dimension and a Lagrange multiplier is shown; solution problems for n>2 are discussed. Second, extrapolation of the model; they include models from group 1 as "cycle cores"; ways to apply heuristics are discussed. Third, non-separable NP tasks that take into account the procedures for vessels entering the simultaneous processing norm. The direction of research and design of the planning model is indicated.
