|
|
|
|
MOR Journal Abstracts
|
 | |
| |
| |
|
Military Operations Research: Responding to Change (LtCol Paul F. Auclair, Edward F. Mykytka and Dr. Greg Parnell)
For the 61st Military Operations Research Symposium, the program chair, Dan Barker, developed a survey to assess how the military operations research (MOR) community was responding to change. The survey was mailed in late 1992 (three years after the fall of the Berlin Wall) and over 1200 of the 3500 surveys were returned. After the Symposium, Paul Auclair, Edward Mykytka and Greg Parnell analyzed the survey results for the journal. The article provides a descriptive summary of the membership, assess the need for professional education, and identifies key insights about how the focus of the community had changed by late 1992. (Pg. 9)
Benchmarking and Efficient Program Delivery for The Department of Defense's Business-Like Activities (Dr. William F. Bowlin)
In recent years, the concept of total quality management (TQM) has been infused into many government operations. A key component of TQM is benchmarking. William F. (Bud) Bowlin illustrates an internal benchmarking approach using a methodology called data envelopment analysis (DEA) and Department of Defense commissary data. By using DEA, an analyst is able to identify superior performance, quantify performance gaps, and contribute to continuous improvement. (Pg. 19)
Bomb Damage Assessment and Sortie Requirements (Dennis K. Evans)
Computer programs that simulate warfare generally do not model bomb damage assessment (BDA) in a complex fashion. Most models either assume perfect BDA or no BDA, or allow the user to select between these two options. Assuming perfect BDA will allow a given force to destroy a given target set much more rapidly than will be the case if there is no BDA. It would superficially appear that perfect BDA and no BDA are opposite limiting extremes. This is not the case. The opposite limiting extremes are actually perfect BDA and extremely bad BDA. "No BDA" is an intermediate case. Hence, it is probably more reasonable to assume no BDA than perfect BDA in doing computer modeling. It is easy to simulate perfect BDA or nonexistent BDA, but any method of modeling imperfect BDA will be highly dependent on the scenario assumed. This suggests that "no BDA" may be preferable to "flawed BDA" as an assumption in computer modeling. (Pg. 31)
Modeling Cost and Schedule Uncertainties — A Work Breakdown Structure Perspective (Paul R. Garvey)
When cost and schedule uncertainty analyses are presented to decision-makers, questions asked with increased urgency are: What is the likelihood of achieving cost and schedule? What is the chance of exceeding the most likely cost for a given schedule? How are cost reserve recommendations driven by schedule uncertainties? Questions such as these are not readily answered by current analysis tools and techniques. This paper advances the state-of-the-practice by describing a family of, less known, multivariate probability models suitable for addressing these and related questions. The application of these models is presented from a work breakdown structure perspective, the traditional framework for developing program resource estimates. (Pg. 37)
Dynamical Instability in Combat Models: Computer Arithmetic and Mathematical Models of Attrition and Reinforcement (Dr. Julian I. Palmore)
Foremost, this paper reports a methodology that can be used to assess information gain and the value of information in combat operations. Secondly, the experiments discussed in this report give modest insight into underlying relationships between tactical intelligence information and combat results. Finally, the paper reports a unique application of information theory to measure intelligence information concerning the size and disposition of enemy forces.
Combat models, nonlinear deterministic models of decision making processes, that deal with attrition of opposing forces, may contain dynamical instabilities and structural variance. Chaos in computation is one cause of instabilities in computer simulations of combat. Another cause of instabilities in decision processes is attributed to timing problems that arise when thresholds are crossed. An example of computer arithmetic effects for Patriot missile software is given in which very small timing errors accumulate and cause gross errors in detection and ranging. Several simple mathematical models of attrition and reinforcement are given and analyzed for chaotic behavior and nonlinear effects. Nonmonotonicity is demonstrated in the response of battle outcomes to changes in resources. (Pg. 45)
Military Training Resource Scheduling: System Model, Optimal and Heuristic Decision Processes (LTC Mike McGinnis, Professor Emmanuel Fernandez-Gaucherand and Dr. Pitu B. Mirchandani)
The United States Army trains thousands of new soldiers each year to fill vacancies in Army organizations. Initial entry training consists of two sequential phases: Basic Combat Training followed by Advanced Individual Training. Until recently, manual heuristic methods were used to schedule hundreds of training companies for initial entry training. We present two approaches for scheduling training resources for the Basic Combat Training phase of initial training. One is a dynamic programming decision model for optimally scheduling training resources. The second is a heuristic procedure implemented in an decision support system (DSS). Computational results are given for both approaches. The heuristic provides "good" solutions in terms of three performance measures: training quality as measured by an instructor-to-trainee ratio, resource utilization, and training costs. (Pg. 53)
Modeling Presentations at National Conventions (Dean S. Hartley III)
Organizing the presentations at professional meetings society meetings is a chore that is common to all professions. The success of general purpose/national conventions is basically measured by the number of presentations that are obtained. Obtaining a maximum number of presentations with a minimum amount of work is a problem that lends itself to analysis using Operations Research. Dean Hartley analyzed the Military Applications Society contributions to national conventions and modeled them using Markov process concepts. The results are useful in minimizing the costs of recruiting session chairs for future meetings. The description of the analysis may also provide a useful education tool. (Pg. 9)
Statistical Validation of a Communications Network Simulation (Ann E.M. Brodeen and Malcolm S. Taylor)
Battlefield communications networks must deliver critical information when and where it is needed despite a rapidly changing and often hostile environment. Reliance upon computer simulations for system development and evaluation is often necessary since most communications systems are to complex to model analytically. Assurance that the simulation model faithfully emulates the process under study is essential in order to establish credibility and support the value of analyses and decisions that may follow. This paper describes a statistical procedure that provides an impartial assessment of agreement between simulated predictions and empirical observations for a communications network. The method is easy to understand, simple to implement, and flexible enough to hold the promise of more general and widespread application. (Pg. 21)
A Data Analysis of Success in OCS, the Use of ASVAB Waivers, and Race (R.R. Read and L.R. Whitaker)
The paper takes an in-depth look at the controversy posed by the facts that the use of ASVAB score waivers for admission to Officer Candidate School (OCS) appears: 1) unrelated to success in OCS when viewed by the individual races; and 2) related to success in OCS when the data are pooled into a single macro set. The short answer is found in the fact that the use of waivers is quite variable from race to race. Further, increasing use of the waiver is associated with decreasing success rates in OCS. It is noted that the use of waivers diminished during the period of the study. The general result would be spurious if the OCS has some sort of racial bias internal to it. Another explanation is that the ASVAB or the administration of the decision rules has a bias that accepts candidates by race group, leading to uneven success rates in the school. (Pg. 29)
Non-Monotonicity, Chaos and Combat Models (J.A. Dewar, J.J. Gillogly, and M.L. Juncosa)
While few combat modelers claim absolute predictivity for their models, many suggest that their models are good at relative prediction — indicating when one system or configuration is better than another. Relatively predictivity requires a model to be monotonic in its outcomes — each additional increment of combat power for one combatant must lead to at least as good an outcome. This report shows that nonlinearities in a very simple deterministic combat models can produce non-monotonic results, where an additional increment of combat power leads to worse results. It further relates these non-monotonicities to the chaos that can bedevil nonlinear systems. (Pg. 37)
Final-Cost Estimates for Research & Development Programs Conditioned on Realized Costs (Mark Gallagher and David Lee)
Managers and their analysts must estimate program costs and completion times. Most R&D programs, however, historically experienced significant schedule slips and incurred dramatic cost increases. Therefore, senior management wants risk assessment of on-going R&D programs. Gallagher and Lee propose a method that presents the probability of various final costs and completion times for an on-going R&D program. Since the approach relies on incurred costs, it avoids the problem of measuring against optimistic budget and schedule projections. Furthermore, while common methods only provide a point estimate, the proposed technique presents the likelihood the final program cost and completion time will be in any particular range. Program managers and their supervisors can use this information to assess the risk in continuing an R&D program. (Pg. 51)
Manoeuvre Warfare: Some Conditions Associated With Success at the Operational Level (D. Rowland, L. R. Speight and M. C. Keys)
This paper is concerned with the conditions likely to enhance the chances of achieving a breakthrough in modern manoeuvre warfare, and then of achieving success in the subsequent campaign. The study which it describes was based on some 159 battles, which a military historian assessed in terms of the success criteria just mentioned, and also rated in terms of some 32 qualitative factors. Of these factors it seems that the achievement of surprise; the employment of aggressive ground reconnaissance; and the condition of air superiority enjoyed by the attacking side, were most strongly associated with success. The study also emphasized the positive relationships which existed between a high rate of terrain capture, a favourable loss ratio and the probability of a favourable campaign outcome. The paper concludes with some discussion of these results, and draws out some implications for the future of battle modelling at the operational level. (Pg. 5)
Exploring a Relationship Between Tactical Intelligence and Battle Results (MAJ E. Todd Sherrill and Dr. Donald R. Barr)
Provided that current doctrine and organization remain the same for a Battalion Task Force, to what degree does increased intelligence information translate into successful land combat operations? Results that we report here suggest a surprisingly modest gain. Of particular interest is the relative small increase in measurable results observed between operations conducted with intelligence products that units can currently expect to have and those operations conducted with "perfect" intelligence information. Additionally, this paper proposes a new MOE for measuring how much information one has about his opponent. The experiment and the methodology have received much acclaim from the joint Staff, J6 that doctrine and organization must co-evolve with technology in order to receive order of magnitude gains in effectiveness and that additional research using our methodology would be very helpful. (Pg. 17)
Illness Incidence During Military Operations as a Soft Operations Research Factor (Christopher G. Blood)
Medical and manpower planners must factor in the impact of unit attrition on the resource requirements of combat operations. While wounded-in-action (WIA) and killed-in-action (KIA) incidence are the most conspicuous examples of battlefield attrition, the incidence of disease and nonbattle injuries (DNBI) also represents a substantial source of personnel noneffectiveness. For operations of moderate to high battle intensity, a significant statistical relationship was observed between casualty rates and DNBI incidence. The dynamics of battlefield DNBI attrition was postulated as the confluence of naturally occurring (peacetime) DNBI incidence with increases in illness incidence resulting from battle fatigue and a weakened immunological system's ability to fend off disease. (Pg. 35)
Realtime Learning of Doctrine and Tactics Using Neural Networks and Combat Simulations (Dr. John D. Morrison)
Limitations of the traditional Artificial Intelligence paradigm restrict its capacity to support manageable and verifiable knowledge base development for expert system simulations. This report argues that because expertise acquired in dynamic military domains is associated with unique aspects of memory and action-response sequences that are resistant to word-based cues and expression, an alternative model is required for acquiring and representing knowledge in these competitive environments.
Motivated by an emerging research into adaptive and neural models, this report documents a USA TRADOC supported research program that proposed and evaluated an adaptive model within the Army's high-resolution combat simulation CASTFOREM. The prototype was designed to support a synthetic model of intelligence that represents complex goal functions, rule-based (deductive) reasoning in the presence of environmental activity that is consistent with expectation, as well as goal-based (inductive) reasoning in the presence of uncertainty — unfamiliar patterns of activity.
The experiment demonstrated that the prototype is not only capable of generating effective tactics, but the prototype converges to stable, rule-based behavior quickly and efficiently. These results motivate further research into the application of intelligent simulations to broader, long-term goals such as developing and optimizing tactics for developmental hardware and software systems. (Pg. 45)
Locational Analyses of Military Intelligence Ground Facilities (A. Mehrez, M. Eben-Chaime and J. Brimberg)
This paper has many contributions. First, an optimization model is developed for the design of ground intelligence systems, a routine task in military intelligence. The model turned out to be new in the location literature, differing from the classical p-median, maximal cover, set covering problems, etc. Second, the integrality of most model variables, an attribute that is most significant in terms of computational tractability, is shown analytically. The design of ground intelligence system is not a trivial task and design quality might be critical to the success of the intelligence operation. The model presented in the paper constitutes an effective and efficient tool that can significantly enhance design quality. Third, insight are gained through computational experiments, including some non-intuitive results such as the large target coverage that can be obtained by very few facilities. Finally, applicative extensions are discussed in details. (Pg. 61)
Scheduling of Military Vehicles Through the Deliberate Nuclear, Biological, and Chemical Decontamination Process (Georgia-Ann Klutke and Valentin Novikov)
The risk to military personnel by the threatened use of chemical weapons during Desert Storm operations in 1992 has led to a reexamination of existing directives for decontaminating exposed troops and equipment. In studying the logistical aspects of the decontamination process, the authors developed a stochastic scheduling model of nuclear, biological, and chemical decontamination activities. An analysis of the model suggests improved strategies for sequencing equipment for decontamination in the field. The scheduling algorithm developed in this work is now a candidate for adoption in ANBACIS, the U.S. Army's decontamination decision support system. (Pg. 71)
Constrained System Optimization and Capability Based Analysis (R. Garrison Harvey, Kenneth W. Bauer Jr. and Joseph R. Litko)
In recent years, military analysts have witnessed the increasing use of response surface methods as applied to analysis of large-scale simulation models. Typically, the output of such an analysis is a stochastic response surface. Often this surface is used for subsequent study, perhaps as an objective function in an optimization problem. This paper presents an alternative to the traditional approaches to optimizing a stochastic response surface subject to constraints. It also presents an extension that offers a new way of quickly visualizing complex relationships among variables in a system. First, the paper focuses on the stochastic nature of the response surface and the implications for subsequent optimization and sensitivity analysis. Next, we present a three step process to evaluate and optimize complex systems. Finally, we present a current application: a unique interactive computer model called Capability Based Analysis (CBA). (Pg. 5)
Ardennes Campaign Simulation (ARCAS) (Walter J. Bauman)
The Ardennes Campaign Simulation (ARCAS) study was performed to improve the credibility of the Stochastic Concepts Evaluation Model (STOCEM), a theater level combat simulation, by comparing the outcome of a STOCEM simulation of the World War II (WW II) Ardennes campaign of 1944-45 (also known as the Battle of the Bulge) with historical campaign results.
Historical campaign data had been developed, using archival sources, into a computerized data base denoted as the Ardennes Campaign Simulation Data Base (ACSDB). The initial positions, configuration, strengths, compositions and availability of forces for the campaign, as depicted in the ACSDB, are used to define the STOCEM force laydown. Representative simulation results (front line movement, major system losses, and personnel casualties) are compared with historical results from the ACSDB. Stochastic variability of model results is also quantified in terms of confidence limits about each sample mean and bounds on sample outcomes. The comparison of simulation results with history is used to develop guidelines for investigating algorithmic changes which may improve model credibility of the STOCEM. Insights on model verification and validation (V & V) are also developed.
Study results indicate that ARCAS STOCEM tends to generate more force movement, weapon system losses, and personnel casualties than occurred in history. Investigations of potential changes to STOCEM logic/inputs suggested by the simulation/history comparisons include simulation of a "breakthrough" attack posture, moderation of attacker move rate in response to a sustained rapid combat advance, and reduction of base lethality against armor for an attacker possessing a high strength advantage. (Pg. 21)
Finding an Optimal Stationing Policy for the US Army in Europe After the Force Drawdown (Andrew G. Loerch, Natashia Boland, Ellis L. Johnson and George L. Nemhauser)
During the Cold War, the United States Army maintained a force of two corps, about 225,000 soldiers, in Europe to deter aggression by the Soviet Union. Following the implementation of the Conventional Armed Forces Europe (CFE) Agreement in 1990 and the dissolution of the Soviet Union in 1991, the force in Europe was gradually reduced to 65,000. In a cooperative effort between analysts at the US Army Concepts Analysis Agency and those at the Computational Optimization Center at the Georgia Tech, Andy Loerch, Natashia Boland, Ellis Johnson, and George Nernhauser developed a large-scale binary integer programming model to assign the US Army units remaining in Europe to installations in an economical manner, and to make recommendations regarding which installations should be candidates for deactivation and closure. (Pg. 39)
Consolidating the USAF's Conventional Munitions Models (Major Kirk A. Yost)
Starting in 1990, the USAF's shrinking budgets caused increased competition for procurement dollars for stocks of conventional munitions. This situation also increased scrutiny of the models used to compute requirements for conventional munitions. Unfortunately, at that time the USAF was supporting four different optimization models to accomplish this job, with the predictable result that different organizations using different models were generating different requirements. In 1995, three of the four models were consolidated into one optimization system, with the aims of using the best ideas in the existing models leveraging investment in common databases, and providing a common baseline for munitions analyses in the USAF. This paper documents the formulation and development of the new Conventional Forces Assessment Model (CFAM) from both a functional and analytical point of view. More importantly, this paper shows it is both analytically and organizationally possible to consolidate existing models and reduce the support requirements while giving the users a better analysis tool. (Pg. 53)
Satisfying Advanced Degree Requirements For US Air Force Officers (Dennis C. Dietz)
In the aftermath of the Cold War, the Air Force has undergone significant organizational changes and force reductions. These actions motivated a major re-engineering study of graduate education operations the Air Force Institute of Technology (AFIT). Lt Col Dietz describes an analytical model that was developed to determine the minimum number of officers that must enter MS or PhD programs each year in order to consistently satisfy all personnel requirements (by academic specialty, degree level, and grade) at minimum cost. The model is formulated as a Markov decision process and solved using linear programming. The model results formed the basis for a specific plan to modify the AFIT personnel structure. Annual cost savings of at least $2.4 million are anticipated. (Pg. 73)
Some
of our links are in portable document format (pdf); they will be indicated by
this symbol
|
