Title: Using Multi Criteria Decision Analysis for Requirements Definition of Systems Architectures
Authors: Donald Bridges, Paul Blessner, Ph.D., Bill Olson, Ph.D.
Abstract: In 2015 the Government Accountability Office (GAO) reported “Defense Acquisition Process: Military Service Chiefs’ concerns reflect need to better define requirements before programs start” (GAO-15-469), where, in their view, the service chiefs collectively expressed dissatisfaction with acquisition program outcomes and believed that the Department of Defense's (DoD) requirements development and acquisition processes needs to be better integrated. The discussion points out that requirements volatility (“scope creep”) due to vagueness in the requirements definition process is a prime source of requirements growth, and that Systems Engineering plays a key role in developing solid requirements.
Too often stakeholders, or requirements definers, rely on historical preference, trade shows, and current system upgrades versus identifying requirements from actual needs assessments. In the high operational tempo communities such as, Naval Special Warfare, FEMA, etc., which require updated technologies to be fielded rapidly, policy makers have a need to analyze approaches that aggregate all available quantifiable data with more subjective knowledge of operators and experts to support decision making. Valid input to the decision making process is key to eliminating waste in the system life cycle. The task with a large number of variables for complex systems is often overwhelming to the requirements developers, therefore, a mathematical decision-based approach, can aide in the weighting of the multitude of variables that are used in existing development models. How we feed the requirements definition machine is key to a successful and useful output.
Decision Theory algorithms have been used for decades in the systems engineering and requirements definition process (Geordadis 2013, Mairiza 2014, Grenn 2015). This report will compare Multi Criteria Decision Analysis family of algorithms for functionality in requirements definition processes, such as weighting criteria for DoDAF and HoQ structures using MATLAB and/or Excel models. A case study using requirements data, source TBD, will be assessed comparing MCDA algorithms (most probably Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), Measuring Attractiveness by a Categorical Based Evaluation Technique (MACBETH) and Analytic Hierarchy Process (AHP)) to evaluate if formulas assist in developing complex requirements definition. A framework for applying algorithms from the MCDA set will also be introduced. Possible future research could be applying these Decision Theory or MCDA, approaches for developing interface standards for combatant craft technology and equipment which contain very complex technical variables and a wide scope of operator needs.