Efficient multinomial selection in simulation

Consider a simulation experiment consisting of v independent vector replications across k systems, where in any given replication one system is selected as the best performer (i.e., it wins). Each system has an unknown constant probability of winning in any replication and the numbers of wins for the individual systems follow a multinomial distribution. The classical multinomial selection procedure of Bechhofer, Elmaghraby, and Morse (Procedure BEM) prescribes a minimum number of replications, denoted as v*, so that the probability of correctly selecting the true best system (PCS) meets or exceeds a prespecified probability. Assuming that larger is better, Procedure BEM selects as best the system having the largest value of the performance measure in more replications than any other system. We use these same v* replications across k systems to form (v*)^k pseudoreplications that contain one observation from each system, and develop Procedure AVC (All Vector Comparisons) to achieve a higher PCS than with Procedure BEM. For specific small-sample cases and via a large-sample approximation we show that the PCS with Procedure AVC exceeds the PCS with Procedure BEM. We also show that with Procedure AVC we achieve a given PCS with a smaller v than the v* required with Procedure BEM. © 1998 John Wiley & Sons, Inc. Naval Research Logistics 45: 459–482, 1998     

Printed circuit board family grouping and component allocation for a multimachine,open-shop assembly cell

This article considers a particular printed circuit board (PCB) assembly system employing surface mount technology. Multiple, identical automatic placement machines, a variety of board types, and a large number of component types characterize the environment studied. The problem addressed is that of minimizing the makespan for assembling a batch of boards with a secondary objective of reducing the mean flow time. The approach adopted is that of grouping boards into production families, allocating component types to placement machines for each family, dividing of families into board groups with similar processing times, and the scheduling of groups. A complete setup is incurred only when changing over between board families. For the environment studied, precedence constraints on the order of component placement do not exist, and placement times are independent of feeder location. Heuristic solution procedures are proposed to create board subfamilies (groups) for which the component mounting times are nearly identical within a subfamily. Assignment of the same component type to multiple machines is avoided. The procedures use results from the theory of open-shop scheduling and parallel processor scheduling to sequence boards on machines. Note that we do not impose an open-shop environment but rather model the problem in the context of an open shop, because the order of component mountings is immaterial. Three procedures are proposed for allocating components to machines and subsequently scheduling boards on the machines. The first two procedures assign components to machines to balance total work load. For scheduling purposes, the first method groups boards into subfamilies to adhere to the assumptions of the open-shop model, and the second procedure assumes that each board is a subfamily and these are scheduled in order of shortest total processing time. The third procedure starts by forming board subfamilies based on total component similarity and then assigns components to validate the open-shop model. We compare the performance of the three procedures using estimated daily, two-day, and weekly production requirements by averaging quarterly production data for an actual cell consisting of five decoupled machines. © 1994 John Wiley & Sons, Inc.     

Minimax resource allocation problems with ordering constraints

Resource allocation problems consider the allocation of limited resources among numerous competing activities. We address an allocation problem with multiple knapsack resource constraints. The activities are grouped into disjoint sets. Ordering constraints are imposed on the activities within each set, so that the level of one activity cannot exceed the level of another activity in the same set. The objective function is of the minimax type and each performance function is a nonlinear, strictly decreasing and continuous function of a single variable. Applications for such resource allocation problems are found, for example, in high-tech industries confronted with large-scale and complex production planning problems. We present two algorithms to solve the allocation problem with ordering constraints. The first one uses characterization of the optimal decision variables to apply a search method. The second algorithm solves a sequence of problems, each in the format of the original problem without ordering constraints. Whereas the computational effort of the first algorithm depends on the desired degree of accuracy even for linear performance functions, the effort of the latter algorithm is polynomial for certain classes of performance functions. © 1994 John Wiley & Sons, Inc.     

Scheduling about a large common due date with tolerance to minimize mean absolute deviation of completion times

In this article the problem of minimizing the mean absolute deviation (MAD) of job completion times about an unrestricted given common due date with tolerance in the n-job, single-machine scheduling environment is considered. We describe some optimality conditions and show that the unrestricted version of the MAD problem with an arbitrary due date tolerance is polynomial by proposing a polynomial-time algorithm for identifying an optimal schedule. © 1994 John Wiley & Sons, Inc.     

Rank‐Cluster‐and‐Prune: An algorithm for generating clusters in complex set partitioning problems

Clustering problems are often difficult to solve due to nonlinear cost functions and complicating constraints. Set partitioning formulations can help overcome these challenges, but at the cost of a very large number of variables. Therefore, techniques such as delayed column generation must be used to solve these large integer programs. The underlying pricing problem can suffer from the same challenges (non‐linear cost, complicating constraints) as the original problem, however, making a mathematical programming approach intractable. Motivated by a real‐world problem in printed circuit board (PCB) manufacturing, we develop a search‐based algorithm (Rank‐Cluster‐and‐Prune) as an alternative, present computational results for the PCB problem to demonstrate the tractability of our approach, and identify a broader class of clustering problems for which this approach can be used. © 2009 Wiley Periodicals, Inc. Naval Research Logistics 2009     

Security,structural factors and sovereignty: Analysing reactions to Kenya’s decision to close the Dadaab refugee camp complex

Kenya’s decision to close the Dadaab refugee camp complex highlights structural flaws in the international refugee regime. While much attention has been paid to Kenya’s reasoning, less has been given to the reactions of organisations and states. Given the state’s primacy in the international system and uncertainty about refugees, Kenya’s decision is perhaps unsurprising. It is contended that the stakeholders were unprepared because of path dependence and disbelief that Kenya would repatriate the refugees. While stakeholder reactions arguably demonstrate concern for refugees, the international refugee regime remains unquestioned, sustaining revenue streams that may fuel corruption, encourage lengthy encampment and prolong conflict.     

Algorithms for solving the conditional covering problem on paths

Consider the conditional covering problem on an undirected graph, where each node represents a site that must be covered by a facility, and facilities may only be established at these nodes. Each facility can cover all sites that lie within some common covering radius, except the site at which it is located. Although this problem is difficult to solve on general graphs, there exist special structures on which the problem is easily solvable. In this paper, we consider the special case in which the graph is a simple path. For the case in which facility location costs do not vary based on the site, we derive characteristics of the problem that lead to a linear‐time shortest path algorithm for solving the problem. When the facility location costs vary according to the site, we provide a more complex, but still polynomial‐time, dynamic programming algorithm to find the optimal solution. © 2005 Wiley Periodicals, Inc. Naval Research Logistics, 2005.