Composite unimodality

This article defines a class of univariate functions termed composite unimodal, and shows how their minimization admits an effective search procedure, albeit one not as efficient as is Fibonacci search for unimodal functions. An approximate Lagrangian approach to an important real-world logistics problem is seen to yield a surrogate problem whose objective function is composite unimodal. The mathematical form of this objective function is likely to be encountered in solving future real-world problems. © 1993 John Wiley & Sons, Inc.     

Sensitivity curves for effective project management

A key problem in project management is to decide which activities are the most important to manage and how best to manage them. A considerable amount of literature has been devoted to assigning “importance” measures to activities to help with this important task. When activity times are modeled as random variables, these activity importance measures are more complex and difficult to use. A key problem with all existing measures is that they summarize the importance in a single number. The result is that it is difficult for managers to determine a range of times for an activity that might be acceptable or unacceptable. In this paper, we develop sensitivity curves that display the most useful measures of project performance (in terms of schedule) as a function of an activity's time. The structure of the networks allows us to efficiently estimate these curves for all desired activities, all desired time ranges, and all desired measures in a single set of simulation runs. The resulting curves provide insights that are not available when considering summarized measures alone. Chief among these insights is the ability to identify an acceptable range of times for an activity that will not lead to negative scheduling consequences. © 2003 Wiley Periodicals, Inc. Naval Research Logistics 50: 481–497, 2003     

Landmarks in defense literature

Preliminary design of an experimental world-circling spaceship RM001 RAND Corporation, Santa Monica, Calif. (1946)     

Solving a (0, 1) hyperbolic program by branch and bound

A problem in (0, 1) hyperbolic programming is formulated and solved by the use of branch and bound methods. Computational results are presented including a comparison among several branching rules. Heuristic methods for quickly finding relatively good feasible solutions are presented and tested. The problem finds application in the scheduling of common carriers. In the solution of the main problem, a subproblem is identified and solved. A geometric analogue is presented, which allows an interesting interpretation of the subproblem. The subproblem itself finds application in the design of gambles.     

An exact method for finding shortest routes on a sphere,avoiding obstacles

On the surface of a sphere, we take as inputs two points, neither of them contained in any of a number of spherical polygon obstacles, and quickly find the shortest route connecting these two points while avoiding any obstacle. The WetRoute method presented here has been adopted by the US Navy for several applications. © 2016 Wiley Periodicals, Inc. Naval Research Logistics 63: 374–385, 2016     

PART II: CONTINUING TO QUESTION THE RELIABILITY OF NUCLEAR DETERRENCE

Responding to Derrin Culp's critique, the author argues that distinguished nuclear theorists may be wrong because groups of experts have been wrong in the past, that city attacks are central to nuclear deterrence theory because killing civilians en masse is what nuclear weapons do best, and that understanding how effective city attacks would be in war is crucial to understanding how well they would work as threats. Moreover, while it is undeniable that nuclear deterrence works some of the time, this simply is not good enough. Because any failure of nuclear deterrence could end in catastrophic nuclear war, nuclear deterrence must be perfect or almost perfect. This is a very difficult standard to reach.     

INCAPACITATING BIOCHEMICAL WEAPONS

Military interest in incapacitating biochemical weapons has grown in recent years as advances in science and technology have appeared to offer the promise of new “non-lethal” weapons useful for a variety of politically and militarily challenging situations. There is, in fact, a long and unfulfilled history of attempts to develop such weapons. It is clear that advances are opening up a range of possibilities for future biological and chemical weapons more generally. The treaties prohibiting biological and chemical weapons make no distinction between lethal and “non-lethal” weapons—all are equally prohibited. Indeed, a sharp and technically meaningful distinction between lethal and “non-lethal” biological and chemical weapons is beyond the capability of science to make. Thus, interest in incapacitating biochemical weapons, and efforts on the part of various states to develop them, pose a significant challenge to the treaty regimes, to the norms against biological and chemical warfare that they embody, and, ultimately, to the essential protections that they provide. Preventing a new generation of biological and chemical weapons from emerging will take concerted efforts and action at the local, national, and international levels.     

A robust queueing network analyzer based on indices of dispersion

We develop a robust queueing network analyzer algorithm to approximate the steady-state performance of a single-class open queueing network of single-server queues with Markovian routing. The algorithm allows nonrenewal external arrival processes, general service-time distributions and customer feedback. The algorithm is based on a decomposition approximation, where each flow is partially characterized by its rate and a continuous function that measures the stochastic variability over time. This function is a scaled version of the variance-time curve, called the index of dispersion for counts (IDC). The required IDC functions for the external arrival processes can be calculated from the model primitives or estimated from data. Approximations for the IDC functions of the internal flows are calculated by solving a set of linear equations. The theoretical basis is provided by heavy-traffic limits for the flows established in our previous papers. A robust queueing technique is used to generate approximations of the mean steady-state performance at each queue from the IDC of the total arrival flow and the service specification at that queue. The algorithm's effectiveness is supported by extensive simulation studies.