Search

The article presents a list of working papers related to management science. Some of the papers include "A Primal-Dual Surrogate Simplex Algorithm: Comparison with the Criss Cross Method," by R. S. Garfinkel and P. L. Yu, "Optimal Impulsive Trajectory Rendezvous Using Projection-Restoration," by W. A. Gruver, and "Asymptotic Normality of the Extremum of Certain Sample Functions," by P. K. Sen.

In this note we consider the validity of substituting a linear goal for a fractional goal in a mathematical program and provide a simple geometric interpretation that contrasts the essential differences between the two programs. [ABSTRACT FROM AUTHOR]

The article presents comments on the management science paper "On Sequencing n Jobs on One Machine to Minimize the Number of Late Jobs," by William L. Maxwell. The paper focused on an integer programming formulation of a one-machine job shop problem. The author contends that Maxwell made in error in proving the validity of an optimal algorithm by applying cutting plane constraints to the problem. The author explains that the problem cannot be solved through traditional cutting plane procedures. The author provides a mathematical model in an attempt to correct the proof.

This paper develops a new algorithm for increasing the revenue in a dynamic product assortment problem. Then, it identifies the challenges faced by managers in practice and discusses the conditions under which workers follow the algorithm. To do so, I conducted a field experiment with a beverage vending machine business. The experiment shows that, on average, workers are reluctant to follow the algorithmic advice; however, the workers are more willing to conform once their forecasts are integrated into the algorithm. Analyses using nonexperimental variations highlight the importance of taking worker and context heterogeneity into account to maximize the benefit from adopting a new algorithm. Higher worker's regret, sales volatility, and fewer delegations increase the conformity, while they mitigate the effects of integration. Workers avoid high-traffic vending machines and focus on machines with high sales volatility when adopting the algorithm. The effects on the sales are largely similar to the effects on product assortments. The results emphasize the gap between nominal and actual performance of an algorithm and several practical issues to be resolved. This paper was accepted by Matthew Shum, marketing. [ABSTRACT FROM AUTHOR]

Ensuring fairness in algorithmic decision making is a crucial policy issue. Current legislation ensures fairness by barring algorithm designers from using demographic information in their decision making. As a result, to be legally compliant, the algorithms need to ensure equal treatment. However, in many cases, ensuring equal treatment leads to disparate impact particularly when there are differences among groups based on demographic classes. In response, several "fair" machine learning (ML) algorithms that require impact parity (e.g., equal opportunity) at the cost of equal treatment have recently been proposed to adjust for the societal inequalities. Advocates of fair ML propose changing the law to allow the use of protected class-specific decision rules. We show that the proposed fair ML algorithms that require impact parity, while conceptually appealing, can make everyone worse off, including the very class they aim to protect. Compared with the current law, which requires treatment parity, the fair ML algorithms, which require impact parity, limit the benefits of a more accurate algorithm for a firm. As a result, profit maximizing firms could underinvest in learning, that is, improving the accuracy of their machine learning algorithms. We show that the investment in learning decreases when misclassification is costly, which is exactly the case when greater accuracy is otherwise desired. Our paper highlights the importance of considering strategic behavior of stake holders when developing and evaluating fair ML algorithms. Overall, our results indicate that fair ML algorithms that require impact parity, if turned into law, may not be able to deliver some of the anticipated benefits. This paper was accepted by Kartik Hosanagar, information systems. [ABSTRACT FROM AUTHOR]

In the era of big data, an important weapon in a machine learning researcher's arsenal is a scalable support vector machine (SVM) algorithm. Traditional algorithms for learning SVMs scale superlinearly with the training set size, which becomes infeasible quickly for large data sets. In recent years, scalable algorithms have been designed which study the primal or dual formulations of the problem. These often suggest a way to decompose the problem and facilitate development of distributed algorithms. In this paper, we present a distributed algorithm for learning linear SVMs in the primal form for binary classification called the gossip-based subgradient (GADGET) SVM. The algorithm is designed such that it can be executed locally on sites of a distributed system. Each site processes its local homogeneously partitioned data and learns a primal SVM model; it then gossips with random neighbors about the classifier learnt and uses this information to update the model. To learn the model, the SVM optimization problem is solved using several techniques, including a gradient estimation procedure, stochastic gradient descent method, and several variants including minibatches of varying sizes. Our theoretical results indicate that the rate at which the GADGET SVM algorithm converges to the global optimum at each site is dominated by an O (1 λ ) term, where λ measures the degree of convexity of the function at the site. Empirical results suggest that this anytime algorithm—where the quality of results improve gradually as computation time increases—has performance comparable to its centralized, pseudodistributed, and other state-of-the-art gossip-based SVM solvers. It is at least 1.5 times (often several orders of magnitude) faster than other gossip-based SVM solvers known in literature and has a message complexity of O(d) per iteration, where d represents the number of features of the data set. Finally, a large-scale case study is presented wherein the consensus-based SVM algorithm is used to predict failures of advanced mechanical components in a chocolate manufacturing process using more than one million data points. This paper was accepted by J. George Shanthikumar, big data analytics. [ABSTRACT FROM AUTHOR]

The problem of determining continuous flows of minimum cost in a network with convex cost functions is considered. The approach used is that of finding, for any given feasible flow, circuit flows of negative incremental costs. In the main theoretical result of this paper, it is proved that if at each stage, given a feasible nonoptimal flow X, the circuit flow with most negative incremental cost is added to X, linear convergence to the optimal solution will be obtained. In addition, this most negative incremental cost determines an upper bound on the difference in cost between the given feasible solution and the optimal. Based on these concepts, an algorithm, which preserves linear convergence, is presented to determine minimum cost flows in networks with convex costs in the arcs. Results of computer runs made for this algorithm are given. The special case of networks with linear costs is also considered. [ABSTRACT FROM AUTHOR]

The main purpose of this paper is to delineate an algorithm for fractional programming with nonlinear as well as linear terms in the numerator and denominator. The algorithm presented is based on a theorem by Jagannathan [7] concerning the relationship between fractional and parametric programming. This theorem is restated and proved in a somewhat simpler way. Finally, it is shown how the given algorithm can be related to the method of Isbell and Marlow [6] for linear fractional programming and to the quadratic parametric approach by Ritter [10]. The Appendix contains a numerical example. [ABSTRACT FROM AUTHOR]

Commentary is presented for an article published in the July 1960 issue of "Management Science," written by L. V. Kantorovich with an introductory note by T. C. Koopmans. According to the author, in his introduction Koopmans addresses in particular a linear programming problem and a matrix game presented by Kantorovich. Also presented is an interpretation of the article's treatment of decision variables and mathematical optimization. The author notes that the article presents algorithms that have been improved from their contemporary presentations.

Managing perishable inventory systems with positive lead times and finite ordering capacities is important but notoriously difficult in both theory and computation. The optimal control policy is extremely complicated, and no effective heuristic policy has been proposed in the literature. In this paper, we develop an easy-to-compute approximation algorithm for this class of problems and prove that it admits a theoretical worst-case performance guarantee under independent and many commonly used positively correlated demand processes. Our worst-case analysis departs significantly from those in the previous studies, requiring several novel ideas. In particular, we introduce a transient unit-matching rule to dynamically match the supply and demand units, and the notion of associated demand processes that provides the right future demand information to establish the desired results. Our numerical study demonstrates the effectiveness of the proposed algorithm. This paper was accepted by Yinyu Ye, optimization. [ABSTRACT FROM AUTHOR]

This paper investigates a decomposition approach for binary optimization problems with nonlinear objectives and linear constraints. Our methodology relies on the partition of the objective function into separate low-dimensional dynamic programming (DP) models, each of which can be equivalently represented as a shortest-path problem in an underlying state-transition graph. We show that the associated transition graphs can be related by a mixed-integer linear program (MILP) so as to produce exact solutions to the original nonlinear problem. To address DPs with large state spaces, we present a general relaxation mechanism that dynamically aggregates states during the construction of the transition graphs. The resulting MILP provides both lower and upper bounds to the nonlinear function, and it may be embedded in branch-and-bound procedures to find provably optimal solutions. We describe how to specialize our technique for structured objectives (e.g., submodular functions) and consider three problems arising in revenue management, portfolio optimization, and healthcare. Numerical studies indicate that the proposed technique often outperforms state-of-the-art approaches by orders of magnitude in these applications. Data and the online appendix are available at https://doi.org/10.1287/mnsc.2017.2849. This paper was accepted by Yinyu Ye, optimization. [ABSTRACT FROM AUTHOR]

Cold start describes a commonly recognized challenge for online advertising platforms: with limited data, the machine learning system cannot accurately estimate the click-through rates (CTR) of new ads and, in turn, cannot efficiently price these new ads or match them with platform users. Traditional cold start algorithms often focus on improving the learning rates of CTR for new ads to improve short-term revenue, but unsuccessful cold start can prompt advertisers to leave the platform, decreasing the thickness of the ad marketplace. To address these issues, we build a data-driven optimization model that captures the essential trade-off between short-term revenue and long-term market thickness on the platform. Based on duality theory and bandit algorithms, we develop the shadow bidding with learning (SBL) algorithms with a provable regret upper bound of O (T 2 / 3 K 1 / 3 (log T) 1 / 3 d 1 / 2) , where K is the number of ads and d captures the error magnitude of the underlying machine learning oracle for predicting CTR. Our proposed algorithms can be implemented in a real online advertising system with minimal adjustments. To demonstrate this practicality, we have collaborated with a large-scale video-sharing platform, conducting a novel, two-sided randomized field experiment to examine the effectiveness of our SBL algorithm. Our results show that the algorithm increased the cold start success rate by 61.62% while compromising short-term revenue by only 0.717%. Our algorithm has also boosted the platform's overall market thickness by 3.13% and its long-term advertising revenue by (at least) 5.35%. Our study bridges the gap between the theory of bandit algorithms and the practice of cold start in online advertising, highlighting the value of well-designed cold start algorithms for online advertising platforms. This paper was accepted by Gabriel Weintraub, revenue management and market analytics. Supplemental Material: Data and the online appendices are available at https://doi.org/10.1287/mnsc.2022.4550. [ABSTRACT FROM AUTHOR]

We introduce a new optimization model, dubbed the display optimization problem, that captures a common aspect of choice behavior, known as the framing bias. In this setting, the objective is to optimize how distinct items (corresponding to products, web links, ads, etc.) are being displayed to a heterogeneous audience, whose choice preferences are influenced by the relative locations of items. Once items are assigned to vertically differentiated locations, customers consider a subset of the items displayed in the most favorable locations before picking an alternative through multinomial logit choice probabilities. The main contribution of this paper is to derive a polynomial-time approximation scheme for the display optimization problem. Our algorithm is based on an approximate dynamic programming formulation that exploits various structural properties to derive a compact state space representation of provably near-optimal item-to-position assignment decisions. As a byproduct, our results improve on existing constant-factor approximations for closely related models and apply to general distributions over consideration sets. We develop the notion of approximate assortments that may be of independent interest and applicable in additional revenue management settings. Lastly, we conduct extensive numerical studies to validate the proposed modeling approach and algorithm. Experiments on a public hotel booking data set demonstrate the superior predictive accuracy of our choice model vis-à-vis the multinomial logit choice model with location bias, proposed in earlier literature. In synthetic computational experiments, our approximation scheme dominates various benchmarks, including natural heuristics—greedy methods, local search, priority rules—and state-of-the-art algorithms developed for closely related models. This paper was accepted by Yinyu Ye, optimization. [ABSTRACT FROM AUTHOR]

In 1964, Kenneth Lane proposed an algorithm to optimize the production schedule of a single-metal, single-processor open pit mine. For this, he proposed a policy based on varying, over time, the so-called "cutoff grade"—or grade threshold used to determine if extracted material should be ore (processed material) or waste (thrown away). Lane's algorithm had a profound impact on the mining industry. However, though it has been used in multiple commercial software systems and has traditionally been taught to every aspiring mining engineer, it is widely considered a heuristic, and little is known regarding the quality of the solutions it produces. In this paper, we formally study Lane's problem. We show that Lane's algorithm can be viewed as an approximate dynamic programming scheme and that Lane's optimality conditions can be formally derived in two different ways: by considering a variant of the problem where the future value function is linearly approximated or by deriving the optimality conditions of a continuous-time version of the problem. We further show that Lane's algorithm can naturally be extended to this continuous-time version of the problem and that when this algorithm converges, it converges to an optimal solution. Finally, through a reformulation, we show that Lane's original problem can be solved using convex mixed-integer programming. Though hypothetical counterexamples can be constructed, computational experiments prove that Lane's algorithm can produce the optimal solution in every real-world data set tested, thereby lending solid support for its practical application. This paper was accepted by Chung Piaw Teo, optimization. [ABSTRACT FROM AUTHOR]

This paper examines the effect of stochastic task times on the total operating costs of a continuously paced assembly line under the assumption that the line is stopped whenever at least one work station requires more time than allotted. A comprehensive stochastic cost function is integrated into an efficient balancing algorithm to enable an approximately minimum cost balance to be obtained. An experiment was conducted to determine the cost savings resulting from using the stochastic method as compared to two deterministic methods. The stochastic method of this paper produced lower cost balances in most cases studied. [ABSTRACT FROM AUTHOR]

In public policy decision making and in capital planning fractional criterion functions occur. For a given set of desirable target values (goals) τi, this paper develops an algorithm for solving a nonconvex programming problem of the type: Minx∈S Maxi{&phii(fi(x)/gi(x) - τi), i = 1,..., m} where fi are convex functions, gi, are concave functions over the convex subset S of Rn and φi are nondecreasing gauge functions. Here φi(&middot) is the penalty incurred whenever the fractional objective fi/gi deviates from the target value τi, the problem is then to choose an x that minimizes the maximum penalty incurred. [ABSTRACT FROM AUTHOR]

This paper combines two rival preference modeling techniques in a new approach to multi-attribute decision making. Currently existing multi-attribute procedures use either global or local preference modeling. In the global modeling technique, a single preference function is constructed in the large; its maximum is the optimal alternative. In the local procedure, sequential approximations of the preference function are constructed in the small. Each approximation generates a trial solution. Under suitable conditions, each trial solution is preferred to its predecessor, so the trial sequence eventually reaches the optimum. Each technique has advantages and disadvantages; this paper combines the desirable features of both techniques in a new improved method. This new method, called the proxy approach, uses the advantages of one technique to overcome the disadvantages of the other. This paper first develops the theoretical aspects of the proxy approach and then compares it to existing procedures. Finally, the proxy algorithm is applied to a curriculum planning problem and numerous insights are gained. [ABSTRACT FROM AUTHOR]

The motivation for this paper is a problem faced by banks which process large volumes of deposited checks. The checks must be separated by bank number before shipment to the Federal Reserve or other banks. The sorting is usually accomplished using a reader-sorter which reads the magnetic ink characters on the checks and separates them into different "pockets." This paper characterizes the optimal sorting strategy and describes an efficient procedure for finding the optimal solution for problems of the size generally found in practice. The algorithm is based on a two state dynamic programming recursion in which characterization theorems are used to drastically reduce the size of the state space and in which the storage requirements are minimal. The paper includes an analysis of computational experience and describes how the algorithm can be used in a real time environment with deadlines. [ABSTRACT FROM AUTHOR]

Mixture models are versatile tools that are used extensively in many fields, including operations, marketing, and econometrics. The main challenge in estimating mixture models is that the mixing distribution is often unknown, and imposing a priori parametric assumptions can lead to model misspecification issues. In this paper, we propose a new methodology for nonparametric estimation of the mixing distribution of a mixture of logit models. We formulate the likelihood-based estimation problem as a constrained convex program and apply the conditional gradient (also known as Frank–Wolfe) algorithm to solve this convex program. We show that our method iteratively generates the support of the mixing distribution and the mixing proportions. Theoretically, we establish the sublinear convergence rate of our estimator and characterize the structure of the recovered mixing distribution. Empirically, we test our approach on real-world datasets. We show that it outperforms the standard expectation-maximization (EM) benchmark on speed (16 times faster), in-sample fit (up to 24% reduction in the log-likelihood loss), and predictive (average 28% reduction in standard error metrics) and decision accuracies (extracts around 23% more revenue). On synthetic data, we show that our estimator is robust to different ground-truth mixing distributions and can also account for endogeneity. This paper was accepted by Serguei Netessine, operations management. [ABSTRACT FROM AUTHOR]

Hub networks play an important role in many transportation and telecommunications systems. This paper introduces a new model called the hub arc location model. Rather than locate discrete hub facilities, this model locates hub arcs, which have reduced unit flow costs. Four special cases of the general hub arc location model are examined in detail. We provide motivation for the new models, and present examples and optimal solutions, using data for U.S. air passenger traffic. Results are used to compare optimal costs, hub locations, and hub arc locations with corresponding hub median optimal solutions. The results reveal interesting spatial patterns and help identify promising cities and regions for hubs. A companion paper (Campbell et al. 2005) presents integer programming formulations and solution algorithms for the new hub arc problems. It also provides details and computation times for these solution algorithms. [ABSTRACT FROM AUTHOR]

This paper proposes and analyzes a periodic review inventory system with two replenishment modes. Regular orders are placed periodically following a base stock policy on inventory position, and arrive at the stocking location after a deterministic lead time. The location also has the option of placing emergency orders, characterized by a shorter lead time but higher acquisition cost, in case of imminent stock outs. Thus, at some appropriate time in the replenishment cycle, the necessity and size of an emergency order is determined according to a base stock policy on net stock. The timing of the emergency order is such that this order arrives and can be used to satisfy the demand in the time period just before the arrival of a regular order, when the likelihood of a stockout is highest. An approximate cost model is developed which can easily be optimized with respect to the order-up-to parameters. This model is used as the basis for a heuristic algorithm, which leads to solutions that are very close to the exact optimal solutions determined through simulation. It is shown that the proposed system offers substantial cost savings relative to a system without the emergency replenishment option. [ABSTRACT FROM AUTHOR]

Consider the linear complementarity problem given in the system: (1) W = MZ + q (2) W ≥ 0, Z ≥ 0 (I) (3) Z[sup T] W = 0 where, W, Z and q are vectors of dimension n. M is a matrix of order n x n and Z[sup T] is the transpose of Z. Any (Z, W) satisfying (1), (2), and (3) is a complementary feasible solution to system (I). In the literature, a class of matrices is defined such that if M belongs to this class, then existence of a feasible solution to system (I) implies the existence of a complementary feasible solution to system (I) with W = 0. In this paper, a new class of matrices M is developed. It is shown that membership of a matrix M in M is equivalent to the property; for any q existence of a feasible solution to system (I) implies the existence of a complementary feasible solution to system (I) for that q with W = 0. This new class of matrices is not contained in any one of the known classes, namely, copositive plus, positive definite or semidefinite, P-matrices, P-matrices, Z-class, etc. [ABSTRACT FROM AUTHOR]

Zero Inventory (ZI) is a new and powerful concept to improve production efficiency. Underpinning ZI is set up reduction and this paper examines what eventuates when set up costs are cut. The folklore is that set up reduction lowers inventory and costs. We obtain results contrary to that. Reducing set up costs need not reduce inventory, for example. Likewise lowering set up costs need not necessarily lower total production and inventory costs. Furthermore, most investments produce marginally decreasing returns. Set up cost reduction prompts the opposite; it produces marginally increasing (not decreasing) returns. A parametric algorithm that calculates the benefit of set up cost reduction is provided. Most presentations of ZI have been descriptive with very few mathematical models of it yet untaken. By modeling certain ZI concepts mathematically, we scrutinize their validity and sometimes invalidity. [ABSTRACT FROM AUTHOR]

Compound option valuation formulae give rise to the summation of a series of multinormal distribution functions. This paper presents an identity on sums of nested multinormal distributions of arbitrary dimension. We show that this identity generalizes some well-known low order identities for the multinormal distribution. We present three applications of the new identity to contingent claims valuation problems. The first and second applications show that by reducing significantly the number of integrals to be evaluated, faster and more accurate algorithms can be developed for implementing the Geske-Johnson American put valuation formula and the Roll-Geske-Whaley American call formula; the third gives new economic insights into the valuation of disaggregated coupon bonds. [ABSTRACT FROM AUTHOR]

The paradox involved in sequences of Prisoner's Dilemma games is due to the fact that game theoretic definitions of optimality rarely coincide with any natural meaning of the word. Decision makers should incorporate their beliefs and experience into any mathematical analysis of the games. Once this has been done, via subjective probabilites, use of the cooperative move in iterated Prisoner's Dilemma games can often be justified. The paper provides a simple algorithm for determining an optimal strategy, once the decision maker's subjective probabilities have been specified. [ABSTRACT FROM AUTHOR]

This paper extends recent work on finding the stochastic queue median (SQM) on a network. The SQM is the optimal location of a single mobile server, when idle, who travels to customers in response to requests for service. Customer demands are limited to the nodes of the network and their requests for service arrive according to a time homogeneous Poisson process, independently from each node. When a service request finds the server busy with a previous customer, it is entered into an M/G/1 queue that is depleted in a first-in, first-out (FIFO) manner. The SQM is the point on the network at which the sum of mean queueing delay and mean travel time is minimized. In this paper the transportation network is restricted to be a tree. By discovering and exploiting convex properties of the objective function and related functions, an efficient finite-step algorithm is found for locating the SQM on a tree. [ABSTRACT FROM AUTHOR]

In recent years, labour negotiations for seven-day-week organizations such as hospitals, mining companies and chemical industries have seen an increased emphasis on improving shift schedules. This paper gives an exact lower bound on the number of workers required to satisfy such primary contractual commitments as ensuring that each employee receives at least A out of every B weekends off, everyone works exactly five days per week, and no one works more than six consecutive days. The model addressed is general in that it allows the number of workers required each day to vary. The proof that the bound is the best possible is constructive. A linear time algorithm is presented that generates schedules satisfying all the primary objectives. Extensions discuss means of incorporating secondary objectives such as giving pairs of adjacent days off. [ABSTRACT FROM AUTHOR]

In this paper, we deal with the two-product, single-machine, static demand, infinite horizon lot scheduling problem. After a quick review of the literature, the necessary and sufficient conditions for feasibility are derived without imposing any preliminary requirements. Then, an optimal solution algorithm is presented. Finally, we show that a necessary condition for the feasibility of N-product schedules is that each cycle time be an integer multiple of some basic cycle time. [ABSTRACT FROM AUTHOR]

This paper describes a hybrid method for solving the multi-commodity transportation and transshipment spatial equilibrium models. The method is basically a specialization of the block successive overrelaxation method for solving a linear complementarity problem with certain block structure and consists of solving a sequence of subproblems of the single-commodity type. These subproblems are solved by a special-purpose principal pivoting algorithm developed in an earlier paper. Under some mild conditions, convergence of the proposed method is established. Finally, computational experience of solving some fairly large randomly generated problems by the proposed hybrid method is presented. [ABSTRACT FROM AUTHOR]

New lock box solution techniques have recently been suggested by Stone and Nauss-Markland. This paper briefly discusses these new methodologies and shows how the algorithms can be combined into a third solution procedure which exploits the computational efficiency of the Stone heuristic and retains the optimizing property of the Nauss-Markland algorithm. [ABSTRACT FROM AUTHOR]

The subject of this paper is a classification and discussion of algorithms for solution of nonlinear pure integer programming problems. The survey is organized by characterizing the mathematical form of the nonlinear optimization problems addressed by the various algorithms. If any method can be used without changes to solve mixed integer nonlinear programming problems that fact is mentioned in the description of the algorithm. However algorithms which must be applied only to mixed integer problems are not surveyed. [ABSTRACT FROM AUTHOR]

In this paper we consider linear multiple objective programs with coefficients of the criteria given by intervals. This class of problems is of practical interest since in many instances it is difficult to determine precisely the coefficients of the objective functions. A subproblem to test if a feasible extreme point is efficient in the problem considered is obtained. A branch and bound algorithm to solve the subproblem as well as computational results are provided. Extensions are discussed. [ABSTRACT FROM AUTHOR]

This paper considers optimization problems with a nonlinear-additive objective function and a single linear constraint. Such models have numerous direct applications and serve as subproblems in procedures for more complex problems. Some important portfolio selection problems can be expressed in this form, and the problem also arises in economic theory. Several authors have noticed independently that special cases and variants of the problem can be solved exactly by surprisingly simple, finite algorithms. The major purpose of this paper is to present these results in a unified framework, which then permits substantial generalizations and extensions. The results lend themselves to an appealing managerial interpretation, similar to the rate-of-return cutoff rules of capital budgeting. [ABSTRACT FROM AUTHOR]

The most popular approach to developing mathematical models of organizational decision processes, the decomposition approach, begins with a mathematical statement of an ideal organizational problem and follows a process of decomposition to derive sub-problems solved by separate units at (possibly) different levels of the organization. Conversely, the composition approach starts with mathematical statements of the subproblems solved by the separate units and proceeds to develop a solution algorithm as a means of coordinating the activities of the separate units. A process of composition must then be followed if one is to discover the derived, organizational problem actually solved. This paper offers an operational definition of an organizational decision process and relates this process to the anatomy of mathematical models purported as being representative of it. Implications regarding the potential of the decision process models as tools of organizational design are explored. [ABSTRACT FROM AUTHOR]

In this paper we study a class of modified policy iteration algorithms for solving Markov decision problems. These correspond to performing policy evaluation by successive approximations. We discuss the relationship of these algorithms to Newton-Kantorovich iteration and demonstrate their covergence. We show that all of these algorithms converge at least as quickly as successive approximations and obtain estimates of their rates of convergence. An analysis of the computational requirements of these algorithms suggests that they may be appropriate for solving problems with either large numbers of actions, large numbers of states, sparse transition matrices, or small discount rates. These algorithms are compared to policy iteration, successive approximations, and Gauss-Seidel methods on large randomly generated test problems. [ABSTRACT FROM AUTHOR]

This paper examines the problem of deriving the tangent (or market) portfolio from a given set of risky assets and a specified risk-free borrowing and lending rate. Deriving the tangent portfolio involves solving a mathematical programming problem which can be specified as the minimization of a quadratic objective function with linear constraints. The complementary pivot algorithm has previously been shown to be capable of deriving the optimal solution to certain quadratic programming problems, subject to a nonnegativity constraint. This paper demonstrates that the algorithm can be used to derive the tangent portfolio and that the nonnegativity constraint does not pose any serious handicap. Furthermore, it is shown that the algorithm can efficiently solve large-scale problems of this nature. [ABSTRACT FROM AUTHOR]

This paper, which is presented in two parts, is a contribution to the theory of fractional programming, i.e. maximization of quotients subject to constraints. In Part 1, duality theory for linear and concave-convex fractional programs is developed and related to recent results by Bector, Craven-Mond, Jagannathan, Sharma-Swarup, et al. Basic duality theorems of linear, quadratic and convex programming are extended. In Part II Dinkelbach's algorithm solving fractional programs is considered. The rate of convergence as well as a priori and a posteriori error estimates are determined. In view of these results the stopping rule of the algorithm is changed. Also the starting rule is modified using duality as introduced in Part I. Furthermore a second algorithm is proposed. In contrast to Dinkelbach's procedure the rate of convergence is still controllable. Error estimates are obtained too. [ABSTRACT FROM AUTHOR]

This paper presents the application of a modified mixed-integer programming algorithm to plan menus in which the portion size of the menu items can vary over a specified positive range. Previous mathematical programming formulations of menu planning problems have either required the variables representing menu items to be bivalent, or have formulated the problem with food groups as decision variables and no integer requirements at all. The former gives rise to a zero-one programming problem and the latter to a "feed-mix" problem. In many instances, a more realistic formulation would require that if a menu item is offered, its portion size must be between a specified upper and lower bound. Although this paper addresses itself chiefly to menu planning, it is readily seen that problems in capital budgeting may be tractable with a similar formulation. It is shown how a branch-and-bound algorithm for mixed-integer programming of the type proposed by Beale and Tomlin can be modified to solve the quasi-integer programming problem resulting from a variable portion size formulation of menu planning problems. [ABSTRACT FROM AUTHOR]

In this paper we discuss the relevancy of the geometric mean as a portfolio selection criteria. A procedure for finding that portfolio with the highest geometric mean when returns on portfolios are lognormally distributed is presented. The development of this algorithm involves a proof that the portfolio with maximum geometric mean lies on the efficient frontier in arithmetic mean variance space. This finding has major implications for the relevancy of much of portfolio and general equilibrium theory. These implications are explored. [ABSTRACT FROM AUTHOR]

This paper presents a systematic procedure for obtaining the optimum packaging frequencies for a number of items which are manufactured jointly but packaged individually immediately after manufacture. The method described in the paper is equally applicable to those problems where the optimum purchasing policy is to be obtained for a number of items procured from a single supplier. An example involving twenty jointly replenished items is solved to illustrate the procedure given in the paper. [ABSTRACT FROM AUTHOR]

This paper investigates the optimal allocation of the marketing budget within the marketing-mix decision variables so that sales (or profit) is maximized in a planning horizon. Since the influence of marketing mix variables upon sales are, in reality, nonlinear and interactive, a geometric programming algorithm is used that solves this problem. A procedure to estimate a functional of sales on the marketing mix and environmental variables utilizing the experienced judgments of the firm's executives and the raw data is provided. The derived functional is later optimized by the Geometric Programming algorithm under a constraint set consisting of budget and strategy restrictions imposed by a firm's marketing environment, and conditions under which the optimal solution is either local or global are identified. An empirical application for a large midwestern brewery is provided which utilizes and illustrates both the estimation an optimization procedures. [ABSTRACT FROM AUTHOR]

We study a new parts-procurement planning problem that is motivated by a global auto manufacturer (GAM) that practices mass customization. Because of the astronomically large number of producible configurations, forecasting their demand is impossible. Instead, firms forecast demand for options that constitute a vehicle. Requirements for many parts (up to 60%) are based on the combinations of options in a fully configured vehicle. The options' forecast, however, does not map into a unique configuration-level forecast. As a result, the options' forecast translates into ranges for many parts' requirements. The combined ranges of a set of parts are not always equal to the sum of the component ranges; they may be less. Determining parts ranges is a large-scale NP-hard problem. Large ranges and inaccurate calculation of these ranges can result in excess inventories, shortages in inventories, and suboptimal flexibility levels. We model and analyze the problem of allocating parts to suppliers and accurately computing the ranges to minimize procurement costs arising because of ranges. The range costs are assumed to be convex increasing. We perform extensive numerical analysis using a large set of randomly generated instances as well as eight industrial instances received from GAM to establish the quality of our approximation framework. Our proposed approach significantly reduces the error in range estimates relative to current industry practice. In addition, the proposed approach for allocations of parts to suppliers reduces joint-parts ranges by an average of 29.87% relative to that of current practice. This paper was accepted by Jeannette Song, operations management. [ABSTRACT FROM AUTHOR]