As the Chief of Naval Operations points out, the demand for automated tools, techniques, and processes is greater now than any time in our history. There are many just reasons for this high demand. We are in a time of great power competition with counter-terrorism operations and significant support to humanitarian operations around the globe all stressing the force. Access to technology at both ends of the spectrum of warfare continues to be easier, and the nefarious intents of individuals and nation states are on display every minute in the news. Our military has great leadership and people, but they struggle to keep up with the pace of operations and change. It is not just operators who are under the gun; program managers, comptrollers, and analysts are overwhelmed by the cacophony of disparate data and policies that flood their systems daily.
Every military service has espoused plans to modernize the plethora of systems acquired over the past few decades in hopes of making better business decisions and reducing the cost of sustainment for legacy applications and programs. This is the mandate for more automation in operations in the field and in program oversight and budgeting in the halls of the Pentagon. But as the adage goes, “If you want it bad, you get it bad.” So, it is time for the practitioners of operations research and analysis to step up to the independent verification and validation plate to ensure that users can trust this new magic. Additionally, with resources being as fragile and uncertain as ever, operations analysis (OA) and operational research (OR) expertise is essential in ensuring that money, time, and talent are applied to the areas where the return on investment is best and most urgently needed.
The federal government is being inundated with new technologies and a plethora of concepts to back them up. Artificial intelligence (AI), quantum computing, big data, blockchain, and the like flood the airwaves daily. What do they all mean? Which technology is the right tool for a specific user? Has the modeling, programming, or formula—what is under the hood—been adequately verified and validated? Are we applying them optimally to solve the most urgent needs or produce the greatest return?
The Solution (or at least part of it)
Just as OR and OA were brought to bear just prior to WWII to address new technology and warfighting, practitioners are now needed to deal with information age and the growing desire to automate more and more of our decision making. Traditional OR and OA models have already been scrutinized, engendering confidence with military decision makers. The OA approach I learned in the Naval Post-Graduate School so long ago continues to be a valid one for approaching many of the problems operators and programmers have today. It is time to apply similar rigor to the wave of new concepts and tools that are coming at the military decision maker.
In the figure above, taken from Skylab Gupta’s Revenue Analytics Online – Machine Learning, Data Science, and Operations Research – What’s the Difference (September 2014), the shifting trend in popularity among the sciences being applied to the management of business is depicted. Gupta’s graphic clearly demonstrates that OR is not as trendy as it once was.
Artificial Intelligence and Data Analytics
Two phrases that are seen and heard in innumerable media outlets are Artificial Intelligence and Data Analytics. These two phrases are common in the media, the halls of the pentagon, and the defense industry. They are interrelated and have more impact on each other than ever before. Has the accuracy with which the internal components capture the operational environment been properly verified and validated? With the demands to do things faster to keep up with the quickening pace of technology, I have my doubts. In the future, the consequences of poorly scrutinizing investments in these areas will have far greater impact than in the past because decision makers are increasingly pressed to blindly believe anything that the system tells them. They have no choice because we have chosen to reduce the number of operators and their skill level to save money.
In her December 2015 blog article, “OR vs. data science vs. analytics: what's in a name?”, Polly Mitchell-Guthrie writes that “views of the relationship between OR and analytics were roughly divided into three camps: OR is a subset of analytics, analytics is a subset of OR, and advanced analytics is the intersection of OR and analytics.” The argument goes on, but the two sciences should coexist to help deal with the growing demand for and more rapid implementation of new concepts.
Multidisciplinary Approaches to Mission Success
We must marry data science to OA and OR practice to ensure that we glean the best answers from our increasing dependence on these new approaches to solving complex problems. OA and OR expertise are called on in development of autonomous systems and in other applications of AI. It necessary that we use of OA and OR in determining where to make the best investments, when to make the investments, and what to stop investing in.
Figure: Data Science Compared With Different Analytics Disciplines – DeZyre blog 20 Oct 2015
The figure above is one of many views on how these new disciplines are connected. It may not be the most accurate picture, but it does show that a large and important cog in the machine is the application of OR.
Vice Chairman of the Joint Chiefs of Staff, General Paul Selva, USAF, recently discussed Pentagon plans to increase investments in AI and man-machine teaming in the 2019 and 2020 budgets, but warned that the Department of Defense is far from achieving the full capability of these technologies. His remarks are in line with the new National Defense Strategy which clearly calls for the US to be more capable and lethal in contested environments and at the higher end of warfare. Advanced autonomous systems, resilient and agile logistics, and big data can help save the future. OR should play a fundamental and critical role in the advancement of these technologies by ensuring that the military buys the right tools, not the slickest thing wrapped in the prettiest box.
Prior to joining LMI, Mr. Harris was a rear admiral with a 34-year career in the United States Navy, where he led joint, combined, multinational, and interagency organizations across all aspects of defense.