Valerdi leads the development of COSYSMO with the support of the Lean Advancement Initiative Consortium Members, USC Center for Systems & Software Engineering Corporate Affiliates, and the Air Force Space & Missile Systems Center.

Valerdi is a member of the NSF-funded IMPACT project that explores the impact of Software Engineering research on programming practice.  Working with Barry Boehm, they are completing a study in the area of software resource estimation.

Valerdi brings together sponsored research projects and a consortium of leaders from industry, government, and academia to answer big-picture questions such as:

  • How do we model complexities in systems?
  • What implications do these complexities have on their life cycle cost?
  • What limitations do humans have in their ability to forecast future events (i.e., cost, schedule, performance of systems)?
  • How can we enable humans so that their decisions are more accurate?

His work is uniquely positioned within the Engineering Systems Division at MIT, as a new kind of interdisciplinary academic unit that spans most departments within the School of Engineering, as well as the School of Science, the School of Humanities, Arts, and Social Sciences, and the Sloan School of Management. His collaborators include colleagues from computer science, management, industrial engineering, aerospace engineering, political science, sociology, and psychology.


The introduction of autonomous systems, and autonomous systems of systems (SoS), have brought significant and interesting challenges. These challenges are not only technological, but also include public perception and operational requirements. As we approach fully autonomous systems over a 30-year horizon, testing becomes enormously more difficult. Therefore, testing and evaluation of unmanned and autonomous systems and SoS has become an area of renewed concern to ensure effective missions while maintaining safety during operation. Current testing approaches lack the ability to predict and adapt to operating environments.

In the complex systems community, alternative approaches to the design of testing can be proposed including solutions based on component technology, design patterns, and resource allocation techniques. It is proposed that an architectural framework for the test and analysis of Unmanned Autonomous Adaptive Systems can provide a key, new contribution: the capability to predict when a system needs to adapt. PATFrame is a prescriptive and adaptive framework for UAS SoS testing to be implemented in a decision support system Prescriptive and Adaptive Testing Framework (PATFrame). The proposed framework is a novel integration of control-theory-based adaptation, multi-criteria decision making and component-based engineering techniques. The goal is to augment traditional DoDAF guidance towards capability-driven development with test architecture and test models specific to the UAS technology sector.