What you should know about approximate dynamic programming

Approximate dynamic programming (ADP) is a broad umbrella for a modeling and algorithmic strategy for solving problems that are sometimes large and complex, and are usually (but not always) stochastic. It is most often presented as a method for overcoming the classic curse of dimensionality that is well‐known to plague the use of Bellman's equation. For many problems, there are actually up to three curses of dimensionality. But the richer message of approximate dynamic programming is learning what to learn, and how to learn it, to make better decisions over time. This article provides a brief review of approximate dynamic programming, without intending to be a complete tutorial. Instead, our goal is to provide a broader perspective of ADP and how it should be approached from the perspective of different problem classes. © 2009 Wiley Periodicals, Inc. Naval Research Logistics 2009     

Building the Somali National Army: Anatomy of a failure, 2008–2018

ABSTRACT

Over a decade of security force assistance (SFA) initiatives to build an effective Somali National Army (SNA) failed because of the interrelated effects of political, contextual and operational challenges. The key political challenges were interest asymmetry between international actors and Somali elites, insufficient focus on institution-building and a lack of donor coordination. The principal contextual challenges in Somalia were the legacies of two decades of state collapse and the negative effects of clan dynamics. The main operational challenges were building an army while simultaneously fighting a war, the complexities of military integration, and the severe capability gaps afflicting the SNA.     

Moving Beyond Population-Centric vs. Enemy-Centric Counterinsurgency

Historically, insurgency is one of the most prevalent forms of armed conflict and it is likely to remain common in the foreseeable future. Recent experiences with counterinsurgency in Iraq and Afghanistan offer many lessons for future counterinsurgents, but the discourse on the subject continues to be mired in a traditional dichotomy pitting population-centric approaches to counterinsurgency against enemy-centric approaches. Historical analysis suggests that this traditional dichotomy is not a sufficiently nuanced way to understand or plan for such operations. Instead, discussions of counterinsurgency should focus on two dimensions: actions (use of physical force vs. political or moral actions) and targets (active insurgents vs. insurgent support). This perspective divides the space of possible counterinsurgency efforts into four quadrants, suggesting that effective counterinsurgency campaigns find a balance of effort across the four quadrants that is well matched to the specific context.     

Technical note: A computationally efficient algorithm for undiscounted Markov decision processes with restricted observations

We present a computationally efficient procedure to determine control policies for an infinite horizon Markov Decision process with restricted observations. The optimal policy for the system with restricted observations is a function of the observation process and not the unobservable states of the system. Thus, the policy is stationary with respect to the partitioned state space. The algorithm we propose addresses the undiscounted average cost case. The algorithm combines a local search with a modified version of Howard's (Dynamic programming and Markov processes, MIT Press, Cambridge, MA, 1960) policy iteration method. We demonstrate empirically that the algorithm finds the optimal deterministic policy for over 96% of the problem instances generated. For large scale problem instances, we demonstrate that the average cost associated with the local optimal policy is lower than the average cost associated with an integer rounded policy produced by the algorithm of Serin and Kulkarni Math Methods Oper Res 61 (2005) 311–328. © 2008 Wiley Periodicals, Inc. Naval Research Logistics 2009     

Precision cyber weapon systems: An important component of a responsible national security strategy?

Given the advances made in conventional weapon capabilities, precision should by now be the accepted and expected norm in cyberspace as well. In this article I argue that developing precision cyber weapon systems, to be used during a lawful conflict, can be an important part of a responsible national security strategy to reduce the amount of violence and physical destruction in conflicts. I first describe a precision cyber weapon system in a military context. I then present three compelling rationales for the development of precision cyber weapon systems based on ethical, operational, and financial considerations. To support the position, I address several arguments that have been levied against their development. Thereafter I present several recommendations for a way ahead.     

Northern Ireland and minimum force: the refutation of a concept?

The debate over whether there exists a British counter-insurgency tradition of minimum force is one that has been thoroughly discussed. Is there anything left to say? This study suggests there is further insight to be attained if one explicitly evaluates the concept of minimum force in relation to the conduct of the security forces in the years of the Northern Ireland conflict. Through an examination of three key periods in the conflict, it will be shown that while there was invariably an awareness of the need to act with restraint among senior officers at the strategic level, this was often difficult to apply at the tactical level in the heat of confrontation. The argument demonstrates that the British Army, and other instruments of the state, rarely acted in a manner that could be described as ‘minimal’. Instead, it was the broader liberal values of the British state that explains largely the degrees of restraint exhibited by the government and security forces.