Estimating the probability that a simulated system will be the best

Consider a stochastic simulation experiment consisting of v independent vector replications consisting of an observation from each of k independent systems. Typical system comparisons are based on mean (long‐run) performance. However, the probability that a system will actually be the best is sometimes more relevant, and can provide a very different perspective than the systems' means. Empirically, we select one system as the best performer (i.e., it wins) on each replication. Each system has an unknown constant probability of winning on any replication and the numbers of wins for the individual systems follow a multinomial distribution. Procedures exist for selecting the system with the largest probability of being the best. This paper addresses the companion problem of estimating the probability that each system will be the best. The maximum likelihood estimators (MLEs) of the multinomial cell probabilities for a set of v vector replications across k systems are well known. We use these same v vector replications to form v^k unique vectors (termed pseudo‐replications) that contain one observation from each system and develop estimators based on AVC (All Vector Comparisons). In other words, we compare every observation from each system with every combination of observations from the remaining systems and note the best performer in each pseudo‐replication. AVC provides lower variance estimators of the probability that each system will be the best than the MLEs. We also derive confidence intervals for the AVC point estimators, present a portion of an extensive empirical evaluation and provide a realistic example. © 2002 Wiley Periodicals, Inc. Naval Research Logistics 49: 341–358, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/nav.10019     

Dominating sets for rectilinear center location problems with polyhedral barriers

In planar location problems with barriers one considers regions which are forbidden for the siting of new facilities as well as for trespassing. These problems are important since they model various actual applications. The resulting mathematical models have a nonconvex objective function and are therefore difficult to tackle using standard methods of location theory even in the case of simple barrier shapes and distance functions. For the case of center objectives with barrier distances obtained from the rectilinear or Manhattan metric, it is shown that the problem can be solved in polynomial time by identifying a dominating set. The resulting genuinely polynomial algorithm can be combined with bound computations which are derived from solving closely connected restricted location and network location problems. © 2002 Wiley Periodicals, Inc. Naval Research Logistics 49: 647–665, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/nav.10038     

An asymptotically optimal greedy heuristic for the multiperiod single‐sourcing problem: The cyclic case

The dynamics of the environment in which supply chains evolve requires that companies frequently redesign their logistics distribution networks. In this paper we address a multiperiod single‐sourcing problem that can be used as a strategic tool for evaluating the costs of logistics network designs in a dynamic environment. The distribution networks that we consider consist of a set of production and storage facilities, and a set of customers who do not hold inventories. The facilities face production capacities, and each customer's demand needs to be delivered by a single facility in each period. We deal with the assignment of customers to facilities, as well as the location, timing, and size of inventories. In addition, to mitigate start and end‐of‐study effects, we view the planning period as a typical future one, which will repeat itself. This leads to a cyclic model, in which starting and ending inventories are equal. Based on an assignment formulation of the problem, we propose a greedy heuristic, and prove that this greedy heuristic is asymptotically feasible and optimal in a probabilistic sense. We illustrate the behavior of the greedy heuristic, as well as some improvements where the greedy heuristic is used as the starting point of a local interchange procedure, on a set of randomly generated test problems. © 2003 Wiley Periodicals, Inc. Naval Research Logistics 50: 412–437, 2003     

Controlling rates in a double queue

A double-ended queue with a Poisson arrival pattern is examined in a situation where the rates depend (in a restricted sense) on both the time and the state of the system. Under some circumstances, the rates can be controlled. This article studies the distribution of the difference in queue sizes for each member of a large class of control strategies and introduces the problem of determining the optimal times at which the control should be in effect in order to maximize certain objective functions.     

A simulation study of the efficiency of empirical Bayes' estimators of multiple correlated probability vectors

Empirical Bayes' methods had been used by Brier, Zacks, and Marlow [1] for estimating performance characteristic vectors of success probabilities. The problem is that of estimating k-dimensional success probabilities of dependent binomial random variables, which are highly correlated. The present study reinforces the results of the previous one by showing, via simulations, that the relative efficiency of the empirical Bayes estimators, compared to the Stein-type and to the maximum-likelihood ones, is very high. This holds even if the success proportions are based on a small number of trials. We study the case of equicorrelation structure with positive correlations.     

An approach for solving discrete alternative multiple criteria problems involving ordinal criteria

The existence of ordinal criteria is a factor that complicates multiple criteria problems. In this article we develop an approach for the problem of choosing among discrete alternatives assuming that criteria are ordinal. The approach requires pairwise comparisons of alternatives by the decision maker. Based on these comparisons, presumably inferior alternatives are eliminated. Experience on randomly generated problems indicates that the approach usually chooses one of the most preferred alternatives and requires a small number of pairwise comparisons.     

Utility assessment procedures for polynomial-exponential functions

This article develops a methodology for testing constant exchange risk properties and identifying an appropriate form for a decision maker's utility function. These risk properties characterize six different utility functions which are sums of products of polynomials and exponential functions. Such functional forms are commonly used in decision analysis applications. The practical advantage of this methodology is that these constant exchange risk properties eliminate the usual arbitrariness in the selection of a parametric utility function and often reduce the data requirements for subsequent estimation. The procedure is straightforward to apply. The decision maker need only provide certainty equivalents for two-outcome gambles and determine the more-preferred gamble in paired comparisons. The technical details of the procedure can be handled by interactive computer software.     

Replacing nonidentical vital components to extend system life

We consider a system that depends on a single vital component. If this component fails, the system life will terminate. If the component is replaced before its failure then the system life may be extended; however, there are only a finite number of spare components. In addition, the lifetimes of these spare components are not necessarily identically distributed. We propose a model for scheduling component replacements so as to maximize the expected system survival. We find the counterintuitive result that when comparing components' general lifetime distributions based on stochastic orderings, not even the strongest ordering provides an a priori guarantee of the optimal sequencing of components. © 2008 Wiley Periodicals, Inc. Naval Research Logistics, 2008     

The Illusion of Afghanistan’s Electoral Representative Democracy: The Cases of Afghan Presidential and National Legislative Elections

This article is the latest in a series of published articles systematically examining Afghan Presidential and legislative elections. Structural problems including fraud, ethno-linguistic block voting, and the Single Non-Transferable Vote have had significant impacts on the development of Afghan democratic elections. The challenge now facing the current Afghan government and future elections is the daunting task of uniting the Afghan people while not repeating the electoral mistakes of the past. The tricky balancing act of fostering an overarching national identity without being perceived as privileging particular identities requires strong leadership and a willingness to challenge traditional ethnic, linguistic, and religious norms when need be. Karzai and Ghani Administrations have seriously failed relative to this dynamic.