Optimal module clustering in program organization

Computer assisted process organization (CAPO) designs a hierarchy chart for a system or program. Published CAPO methods develop a single measure for intermodule connections based on quantification of the most common measures such as cohesion and data coupling. Utilizing common attributes described by a data flow diagram and complete system dictionary, a mathematical clustering model then describes interactions among system or program modules. When used in conjunction with integer programming methods, the clustering model derives a hierarchy chart of organized processes in a top-down, optimal fashion. Extensions to existing CAPO methods include guidelines to allow the model to iterate successive levels of a complete hierarchy and a procedure to determine the number of modules at each level. The model is applied to a sample business application and the resulting design is compared to a manually derived design.<
>