A computer-implemented method for creating optimized sub-projects for a project. Boundary conditions, input regarding output objects, and input regarding “as-is” data sources are received. The output objects are decomposed into data objects and logical processes used to create the output objects. Value clusters are determined. The data objects are organized into “to be” data structures and the “to be” which are mapped to the “as-is” data sources. Additional processes are determined for moving data from a source to a target. An affinity matrix is created based on the value clusters. Finally, an optimization operation is executed with an optimization engine to produce the optimized sub-projects. The optimization engine takes as inputs the above information.