A robust regression technique using compound estimation

Least squares fitting of regression models is a widely used technique. The presence of outliers in the data can have an adverse effect on the method of least squares, resulting in a model that does not adequately fit to the bulk of the data. For this situation, robust regression techniques have been proposed as an improvement to the method of least squares. We propose a robust regression procedure that performs well relative to the current robust methods against a variety of dataset types. Evaluations are performed using datasets without outliers (testing efficiency), with a large percentage of outliers (testing breakdown), and with high leverage outliers (testing bounded influence). The datasets are based on 2-level factorial designs that include axial points to evaluate leverage effects. A Monte Carlo simulation approach is used to evaluate the estimating capability of the proposed procedure relative to several competing methods. We also provide an application to estimating costs for government satellites. © 1998 John Wiley & Sons, Inc. Naval Research Logistics 45: 125–139, 1998     

Cost-based due-date assignment with the use of classical and neural-network approaches

Traditional methods of due-date assignment presented in the literature and used in practice generally assume cost-of-earliness and cost-of-tardiness functions that may bear little resemblance to true costs. For example, practitioners using ordinary least-squares (OLS) regression implicitly minimize a quadratic cost function symmetric about the due date, thereby assigning equal second-order costs to early completion and tardy behavior. In this article the consequences of such assumptions are pointed out, and a cost-based assignment scheme is suggested whereby the cost of early completion may differ in form and/or degree from the cost of tardiness. Two classical approaches (OLS regression and mathematical programming) as well as a neural-network methodology for solving this problem are developed and compared on three hypothetical shops using simulation techniques. It is found for the cases considered that: (a) implicitly ignoring cost-based assignments can be very costly; (b) simpler regression-based rules cited in the literature are very poor cost performers; (c) if the earliness and tardiness cost functions are both linear, linear programming and neural networks are the methodologies of choice; and (d) if the form of the earliness cost function differs from that of the tardiness cost function, neural networks are statistically superior performers. Finally, it is noted that neural networks can be used for a wide range of cost functions, whereas the other methodologies are significantly more restricted. © 1997 John Wiley & Sons, Inc.     

A simple approximation for a multistage capacitated production-inventory system

We develop a simple approximation for multistage production-inventory systems with limited production capacity and variable demands. Each production stage follows a base-stock policy for echelon inventory, constrained by production capacity and the availability of upstream inventory. Our objective is to find base-stock levels that approximately minimize holding and backorder costs. The key step in our procedure approximates the distribution of echelon inventory by a sum of exponentials; the parameters of the exponentials are chosen to match asymptotically exact expressions. The computational requirements of the method are minimal. In a test bed of 72 problems, each with five production stages, the average relative error for our approximate optimization procedure is 1.9%. © 1996 John Wiley & Sons, Inc.     

Artificial intelligence & future warfare: implications for international security

Recent developments in artificial intelligence (AI) suggest that this emerging technology will have a deterministic and potentially transformative influence on military power, strategic competition, and world politics more broadly. After the initial surge of broad speculation in the literature related to AI this article provides some much needed specificity to the debate. It argues that left unchecked the uncertainties and vulnerabilities created by the rapid proliferation and diffusion of AI could become a major potential source of instability and great power strategic rivalry. The article identifies several AI-related innovations and technological developments that will likely have genuine consequences for military applications from a tactical battlefield perspective to the strategic level.     

Optimal policies for stochastic clearing systems with time-dependent delay penalties

We study stochastic clearing systems with a discrete-time Markovian input process, and an output mechanism that intermittently and instantaneously clears the system partially or completely. The decision to clear the system depends on both quantities and delays of outstanding inputs. Clearing the system incurs a fixed cost, and outstanding inputs are charged a delay penalty, which is a general increasing function of the quantities and delays of individual inputs. By recording the quantities and delays of outstanding inputs in a sequence, we model the clearing system as a tree-structured Markov decision process over both a finite and infinite horizon. We show that the optimal clearing policies, under realistic conditions, are of the on-off type or the threshold type. Based on the characterization of the optimal policies, we develop efficient algorithms to compute parameters of the optimal policies for such complex clearing systems for the first time. We conduct a numerical analysis on the impact of the nonlinear delay penalty cost function, the comparison of the optimal policy and the classical hybrid policy (ie, quantity and age thresholds), and the impact of the state of the input process. Our experiments demonstrate that (a) the classical linear approximation of the cost function can lead to significant performance differences; (b) the classical hybrid policy may perform poorly (as compared to the optimal policies); and (c) the consideration of the state of the input process makes significant improvement in system performance.     

Brazilian National Defence Policy: foreign policy,national security,economic growth,and technological innovation

ABSTRACT

With the world’s ninth largest economy and comprising nearly 60% of South America’s GDP, 47% of its territory, and 49% of its population, Brazil has become a regional power and an important actor in world affairs over recent decades. This scenario has led the government to re-evaluate its role in the world order, resulting in the enactment of the National Defence Policy, whose objective was to consolidate the country as a regional power while at the same time addressing national security issues, promoting economic development through a series of defence programmes, restructuring the defence industrial base, fostering innovation through technology and knowledge transfer to Brazil, and indigenous research and development. However, the policy’s implementation suffers from several challenges discussed in this article, which may test the capability and competence of Brazilian policymakers, military, industrialists, and other individuals and organisations involved in its implementation.     

Evaluation of base-stock policies in multiechelon inventory systems with state-dependent demands part I: State-independent policies

This article analyzes a model of a multiechelon inventory system. The exogenous demands form Markov-modulated Poisson processes. That is, the demand rates are functions of an underlying Markov chain. Each location follows a base-stock policy which is independent of the state of the underlying Markov chain. We employ the exogenous transit mechanism introduced by Zipkin [7] and Svoronos and Zipkin [6]. The transit times between locations have phase-type distributions. An exact procedure to compute steady-state performance measures is presented. © 1992 John Wiley & Sons, Inc.     

Scheduling of parallel processors: A posterior bound on LPT sequencing and a two-step algorithm

This article considers the problem of scheduling parallel processors to minimize the makespan. The article makes two key contributions: (1) It develops a new lower bound on the makespan for an optimal schedule, and (2) it proposes an efficient two-step algorithm to find schedules of any desired accuracy, or percent above optimal. In addition, a posterior bound on LPT (longest processing time) sequencing is developed in the article. It is proved that this bound dominates the previously reported bounds on LPT sequencing.