Multilevel-multihorizon control of production system in uncertain conditions

The paper considers production control problem for large-scale systems with continuous elementary production processes. The multilevel-multihorizon method due to Milkiewicz was applied. The method leads to solution of a sequence of optimization subproblems defined for individual time horizons. Each subproblem consists of a production planning subproblem and a production scheduling subproblem, solved sequentially. Uncertainty is mainly treated in the planning subproblems. The subproblems are solved with LP methods. Multilevel control is implemented with a superior coordination system. The basic tasks of this system are: acquisition of data on the functioning of the system and on external conditions, decision making, verification and modification of models. The paper is a continuation of an interesting suggestion (including, for instance, classification of LP problems according to time horizons: deterministic for current problems, stochastic for longer time horizons and fuzzy for the longest time horizons considered), though one would very gladly see a practical implementation or at least an in-depth empirical study of the approach.