Improved conceptual estimating performance using a knowledge-based approach

This dissertation presents the foundations of a knowledge-based approach for improving and measuring conceptual estimating performance. Conceptual estimating is defined as the forecasting of project costs with an incomplete scope definition and before any significant amount of information is available from detailed project design. Because of these information constraints, conceptual estimating needs the contextual application of estimating expertise and judgment. This study addresses the question of how conceptual estimating knowledge and information can be acquired and used to solve specific problems of this domain in order to achieve better performance. It also demonstrates how these inputs can be used in the assessment of the expected quality of conceptual estimates. Knowledge-based Expert Systems (KBES) concepts are applied to develop a conceptual estimating architecture that integrates knowledge and expertise to emulate the performance of an experienced estimator. The study of conceptual estimating literature and a survey on estimating practices reveals that scope definition is the most critical challenge of the estimating problem and should be the real target of estimating efforts. The issues of scope definition are explored and ideas for their resolution proposed. A new scope modeling procedure is designed, and an object-oriented system is selected for scope knowledge representation. A scope modeling system prototype is implemented in a microcomputer, to demonstrate the capabilities of this approach. The problem of assessing the expected accuracy and realiability of estimates is also addressed in this research. Through conceptual analysis an accuracy causal model is developed. This model is utilized to assess estimates' accuracy and reliability using knowledge acquired from experienced estimators. A microcomputer prototype system is also designed and implemented to demonstrate this quality assessment methodology.