When several quantitative variables are related through constraints and objectives, it can be difficult to understand why a certain quantity changes in magnitude after certain changes, or why a certain number seems larger or smaller than expected (as compared to a reference value). Large organizations which seek to optimize very large numbers of parameters to achieve constraints such as supply-demand matching face such problems, where the relationships between variables are controlled by a mixture of human processes and software algorithms. This disclosure describes scalable, flexible frameworks and searching techniques that improve transparency, e.g., enable root-cause analysis, for large families of variables. The techniques enable the understanding of the difference between two data-generating flows (versions) with comparable inputs, intermediate variables, and output variables.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.