A multi-criteria partitioning technique for information system design

Abstract

In order to design and operate a complex computer based information system, one needs to partition its transformation process into modules of manageable size. The same process can be partitioned in many different ways, each alternative partition having a different effect on the system's multiple success criteria. The optimal partition of a given process can be found by a four-step technique: 1. Identify significant interfunctional relationships. 2. Rank system success criteria in sequence of user preference. 3. Apply known partitioning rules in preference sequence to find a set of near-optimal partitions. 4. Select the final part using judgement, prototyping or other techniques.