In the planning of a software development project, managers must estimate the amount of effort needed for distinct phases of activity. A number of rules of thumb exist in the literature to help the program manager in this task. However, very little work has been done to validate these rules of thumb. Applying least square models and Hotelling's$T^{2}$test, we evaluate these rules of thumb against a large database of Department of Defense projects. We determine that variability limits the simple application of any such rule. However, there are some worthy of closer attention, and we recommend adjustments for improved application. We also determine that projects which give extra attention to early phases experience less schedule growth. These findings were robust across developmental process type, military service, and project size.

Traditional genetic programming only supports the use of arithmetic and logical operators on scalar features. The GTMOEP (Georgia Tech Multiple Objective Evolutionary Programming) framework builds upon this by also handling feature vectors, allowing the use of signal processing and machine learning functions as primitives, in addition to the more conventional operators. GTMOEP is a novel method for automated, data-driven algorithm creation, capable of outperforming human derived solutions.