An adaptive approach in production scheduling based on the mixed graph model

Abstract

Proposes computer-aided heuristic approach for production planning and scheduling based on the ideas of adaptation. The objective is to make the scheduling theory more useful for operations management and production research workers. The methodology proposed bridges exact algorithms developed in the frameworks of scheduling theory, and methods of production scheduling. It is devised to combine merits of the enumerative algorithms based on the mixed graph model and heuristic algorithms oriented on the real-life problems. The former ones may, in principle, ensure high quality of the solution via exhausting enumeration. However, it is the case for problems of limited dimension only. The latter ones may be rather efficient since they are developed for practical problems of large dimension taking into account the peculiarities of such a problem. The main idea of the adaptive approach is to produce for the given class of similar problems effective (in sense of good performance measure) and efficient (in computational sense) heuristics by computer-aided adaptation (tuning). It is realized automatically at the learning stage while solving sample problems of the class in consideration using scheduling theory apparatus, accumulating knowledge, simplifying and structuring it. This experience is applied further on for solving problems of the same class at the examination stage.