Application of the linear-quadratic-Gaussian tracking problem to production planning and control systems

The application of the linear-quadratic-Gaussian (LQG) tracking problem to production planning and control systems is discussed. Decision concepts are developed from an examination of the structure of a generalized production control system. The decision problem is modeled as a LQG tracking problem operating in the presence of persistent disturbances--the disturbances representing the demand for finished goods. The development of a solution leads to a linear, time-invariant decision policy which is a decomposition of the decision input into separate and nearly independent decision tasks. Sufficient conditions for the existence of a solution are presented. An open-loop, closed-loop decision system is presented to show how the decision model can be used to analyze relationships among system structure, performance, and information requirements.