Abstract: The paper explores the problem of optimal hierarchical stochastic production planning and control in agile manufacturing workshops (AMW). A stochastic nonlinear programming model of production with delay interaction (which is a static optimization model to solve the dynamic optimization problem) is built up and transformed into a deterministic nonlinear programming model and further into a linear programming model by adding constraints. Then, a Karmarkar's algorithm and an interaction/prediction algorithm are used to solve the model, and the corresponding programs have been written. Through hierarchical stochastic production planning examples, the Karmarkar's algorithm, interaction/prediction algorithm and linear programming method in Matlab are compared with, thus showing that the proposed approaches are very suitable for optimally decomposing AMW's random product demand plans into short-term stochastic plans to be executed by FMS in AMW, especially for the case where workpieces are transferred between FMS through a shop storage with a delay of a production period.