A Decision Support System for planning chemical production of active ingredients in a pharmaceutical company

Chemical production processes in the pharmaceutical industry are often characterized by multiple produced goods at a subset of manufacturing steps. Some of the co- and by-products are recycled and as a consequence the network flow representation of the production process contains cyclical structures. Commercial software packages for MRP type production planning cannot handle those structures appropriately. Therefore a Decision Support System (DSS) has been developed to determine the time phase production quantities of all active ingredients and intermediates for the whole planning horizon. Problems of this kind can be formulated as mixed integer programming models. A dialog system links these models to the data bases containing production data such as bills of material, routings and capacities. It will be shown how design of data bases and dialogs has been performed using as the main tools a relational type of data base (ADABAS) and a fourth generation language (NATURAL). Finally some experience will be communicated concerning both the implementation of the system and its use by the planning department for pharmaceutical production. The acceptance of the DSS has been achieved by strict adherence to two guide-lines: First the DSS must enable the planner to perform at least all of the planning steps and functions he is used to in an MRP II (Manufacturing Resource Planning) environment. Second and presumably more important Operational Research OR solution methodology (such as LP) must be applied in a way that all results are explicable to a user who has no knowledge of quantitative solution techniques except MRP calculation schemes.