Constraint Programming

In the Industrial 4.0 era, much more data is entered into production planning and the selection of suitable production and storage locations than before. In view of the large number of decision-relevant data, it is best to use mathematical methods to calculate which product should be produced at which location and when. In this context, linear programming is of particular importance. The article describes its use by means of practical examples. On IBM Decision Optimization Center as technical platform, software partners can create applications for better decisions in all areas of production and location optimization, including the entire production and logistics network.

The objective of designing layout for a chemical plant is to place all necessary equipments within a steel construction and predefined building sites. The design must consider constraints arising from various areas of expertise. An installation is optimal if as many constraints as possible are satisfied and the costs are minimized. The difficulty in finding an optimal installation lies in the combinatorial complexity of the problem. Moreover the constraints are often expressed in a fuzzy way and may be contradictory. Therefore an optimal installation can often not be achieved. In this article we present two new approaches towards computer aided plant design. The first solution combines a mulitobjective evolutionary algorithm with fuzzy logic. The second solution uses constraint programming.