Hierarchical Production Planning

Abstract

Hierarchical production planning basically involves the decomposition of decision problems in a production environment and the design of a sequential solution procedure, the required information and the organizational systems. A hierarchical decision process in a decentralized organization implies that upper level decisions are made first to impose constraints upon the lower level decisions. It is an important issue to design the linking mechanism between different levels in a manner so as to guarantee consistency and feasibility. This linking mechanism in fact defines the information flow between the levels, and the aggregation/disaggregation procedure of decision variables and production data should also be designed carefully to permit information exchange. Mathematical modelling is frequently employed to characterize the decision process in each level, and the level of detail, the length of planning horizons, the objective functions, the decision variables and the decision-makers usually differ in each model.In this study a hierarchical production planning model is developed for a detergent manufacturing company. A bi-level product hierarchy is designed to support the bi-level scheduling system, where the decision problem in each level is further decomposed into sub-problems. Different mathematical solution techniques including optimization models, Lagrangean relaxation and heuristic dispatching rules are employed to reflect the actual planning process of the company under consideration. Furthermore performance measures are defined for model validation, and some preliminary follow-up results are cited to verify the model adequacy.Key wordsHierarchical production planningaggregation/disaggregationschedulingsequence-dependent setupssub-gradient optimizationtraveling salesman problem