The data-driven, multi-item, one-warehouse, multiretailer (OWMR) problem is examined by leveraging historical data and using machine learning methods to improve the ordering decisions in a two-echelon supply chain. A deep stacking kernel machine (DSKM) and its adaptive reweighting extension (ARW-DSKM), fusing deep learning and support vector machines, are developed for the data-driven, multi-item OWMR problems with backlog and lost sales. Considering the temporal network structure and the constraints connecting the subproblems for each item and each retailer, a Lagrange relaxation–based, trilevel, optimization algorithm and a greedy heuristic with good theoretical properties are developed to train the proposed DSKM and ARW-DSKM at acceptable computational costs. Empirical studies are conducted on two retail data sets, and the performances of the proposed methods and some benchmark methods are compared. The DSKM and the ARW-DSKM obtained the best results among the proposed and benchmark methods for the applications of ordering decisions with and without censored demands and with and without new items. Moreover, the implications in selecting suitable, that is, prediction-then-optimization and joint-prediction-and-optimization, frameworks, models/algorithms, and features are investigated. History: Accepted by Ram Ramesh, Area Editor for Data Science and Machine Learning. Funding: This work was supported by the National Natural Science Foundation of China [Grant 72371062]. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplemental Information ( https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2022.0365 ) as well as from the IJOC GitHub software repository ( https://github.com/INFORMSJoC/2022.0365 ). The complete IJOC Software and Data Repository is available at https://informsjoc.github.io/ .

The classic spiking neural P (SN P) systems abstract the real biological neural network into a simple structure based on graphs, where neurons can only communicate on the plane. This study proposes the hypergraph-based numerical spiking neural membrane (HNSNM) systems with novel repartition protocols. Through the introduction of hypergraphs, the HNSNM systems can characterize the high-order relationships among neurons and extend the traditional neuron structure to high-dimensional nonlinear spaces. The HNSNM systems also abstract two biological mechanisms of synapse creation and pruning, and use plasticity rules with repartition protocols to achieve planar, hierarchical and spatial communications among neurons in hypergraph neuron structures. Through imitating register machines, the Turing universality of the HNSNM systems is proved by using them as number generating and accepting devices. A universal HNSNM system consisting of 41 neurons is constructed to compute arbitrary functions. By solving NP-complete problems using the subset sum problem as an example, the computational efficiency and effectiveness of HNSNM systems are verified.

This study develops machine learning methods for the data-driven demand estimation and assortment planning problem by addressing three subproblems, that is, demand forecasting simultaneously considering cross-selling and substitutions, estimation of the cross-selling and substitution effects, and assortment optimization. These three subproblems are transformed into three sequentially related machine learning problems: collective demand forecasting, demand inference for cross-selling and substitutions, and assortment rule mining. For collective demand forecasting, related product features are introduced to consider both the cross-selling and substitution effects, and a collaborative coordinate descent method with a good convergence property is developed to make distributed demand forecasting and a global update of related product features. Using the results, demand inference adopts transfer and semisupervised learning methods to tackle the challenge of missing data in quantifying the cross-selling and substitution effects. For assortment rule mining, the assortment rules bridge the gap between prediction and optimization, and the developed heuristics obtain the best assortment using the prior knowledge discovered in demand inference. The computational results on a real-world database and a semisynthetic database show that collective demand forecasting obtained far better results than the standard demand forecasting methods and some popular graph learning methods, and the developed heuristics identified much better assortments than those obtained with the baseline methods. History: Accepted by Ram Ramesh, Area Editor for Data Science and Machine Learning. Funding: This work was supported by the construction base project of discipline innovation and talent introduction plan of Chinese higher educational institutions (111 project) [Grant B16009] and the National Natural Science Foundation of China [Grant 72031002]. Supplemental Material: The online appendices are available at https://doi.org/10.1287/ijoc.2022.1251 .

The prevalence of omnichannel marketing and sales enables firms to make ordering decisions based on multisource heterogeneous information from all the channels. This work extends the data-driven inventory management models with single-source or homogeneous information to those with multisource heterogeneous information. Representation learning and information fusion strategies are used to deal with the challenges of high dimensionality and heterogeneity of multisource heterogeneous information. Quantile regression convolutional attention neural networks with different structures embedding different information fusion strategies are developed to address the problems. Two cases with real-world data are studied, and the results show that the proposed methods using representation learning and/or information fusion strategies have (far) better performances than the existing methods without using these strategies. Moreover, some managerial insights into representation learning and information fusion, different from the practices of deep learning in computer vision and natural language processing, are provided for inventory management problems.

This study develops an improved failure mode and effects analysis (FMEA) method using two-dimensional uncertain linguistic variables (2DULVs) and alternative queuing method (AQM). The 2DULVs are employed to represent the evaluations provided by FMEA team members on the weights of risk factors and the risk of failure modes in the form of pairwise comparisons. The two-dimensional uncertain linguistic best worst method (2DUL-BWM) is used to derive the weights of risk factors. The two-dimensional uncertain linguistic AQM (2DUL-AQM) is proposed for determining the risk ranking of the identified failure modes. Finally, the maintenance of a water treatment plant is presented as an example to demonstrate the applicability and effectiveness of the proposed FMEA method. By comparing its performance with those of other existing methods, the proposed FMEA is shown to be more advantageous in ranking the risk of failure modes.

Abstract In many applications of multiple objective network programming (MONP) problems, only integer solutions are acceptable as the final optimal solution. Representative efficient solutions are usually obtained by sampling the efficient set through the solution of augmented weighted Tchebycheff network programs. Because such efficient solutions are usually not integer solutions, a branch‐and‐bound (BB) algorithm is developed to find integer efficient solutions. The purpose of the BB algorithm is to support interactive procedures by generating representative integer efficient solutions. To be computationally efficient, the algorithm takes advantage of the network structure as much as possible. An algorithm, used in the BB algorithm and performed on the key tree, is developed to construct feasible solutions from infeasible solutions and basic solutions from nonbasic solutions when bounds on branching variables change. The BB algorithm finds basic and nonbasic or supported and unsupported integer efficient solutions as long as they are optimal. Details of the algorithm are presented, an example is provided and computational results are reported. Computational results show that the BB algorithm performs well. Although the BB algorithm is developed for the purpose of generating integer efficient solutions for MONP problems, it can also solve more general integer network flow problems with linear side constraints. © 2013 Wiley Periodicals, Inc. Numer Methods Partial Differential Eq 2013

Over 20 studies have been published in the past decade involving file and data type classification for digital forensics and information security applications. Methods using n-grams as inputs have proven the most successful across a wide variety of types; however, there are mixed results regarding the utility of unigrams and bigrams as inputs independently. In this study, we use support vector machines (SVMs) consisting of unigrams and bigrams, as well as complexity and other byte frequency-based measures, as inputs. Using concatenated unigrams and bigrams as input and a linear kernel SVM, we achieve significantly improved results over those previously reported (73.4% classification rate across 38 file and data types). We are the first to use concatenated n-grams as the sole input, and we show their superiority over inputs used previously. We also found that too many different types of features as inputs result in overfitting and poor generalization properties. We include several types seldom or not studied in the past (Microsoft Office 2010 files, file system data, base64, base85, URL encoding, flash video, M4A, MP4, WMV, and JSON records). The “winning” approach is instantiated in an open source software tool called Sceadan - Systematic Classification Engine for Advanced Data ANalysis.

For many practical multiple objective network programming (MONP) problems, only integer solutions are meaningful and acceptable. Representative efficient solutions are usually generated by solving augmented weighted Tchebycheff network programs (AWTNPs), sub-problems derived from MONP problems. However, efficient solutions generated this way are usually not integer valued. In this study, two algorithms are developed to construct integer efficient solutions starting from fractional efficient solutions. One algorithm finds a single integer efficient solution in the neighborhood of the fractional efficient solution. The other enumerates all integer efficient solutions in the same neighborhood. Theory supporting the proposed algorithms is developed. Two detailed examples are presented to demonstrate the algorithms. Computational results are reported. The best integer efficient solution is very close, if not equal, to the integer optimal solution. The CPU time taken to find integer efficient solutions is negligible, when compared with that taken to solve AWTNPs. © 2010 Wiley Periodicals, Inc. NETWORKS, Vol. 57(4), 362-375 2011

Three warm-start routines are developed to find initial basic feasible solutions for augmented weighted Tchebycheff network programs, subproblems derived from multiple-objective network-programming problems. In an interactive solution procedure, a series of augmented weighted Tchebycheff network programs need to be solved sequentially to find representative nondominated solutions. To speed up the solution process using the network structure of the problem, these warm-start routines start the solution process of one augmented weighted Tchebycheff network program from the optimal solution of the previous one. All three warm-start routines use the same strategy but different ways of reducing the number of basic flow variables, or equivalently increasing the number of basic nonflow variables to construct a basic solution. These warm-start routines can be used by any interactive procedures to facilitate the solution process of multiple-objective network-programming problems. A detailed example is presented. A computational experiment is conducted to compare the performance of these warm-start routines. A cold-start routine and NETSIDE, specialized software for solving network problems with side constraints, are also used as references in the experiment. These warm-start routines can save substantial computation time.

MOTIVATION: Extracting useful information from expression levels of thousands of genes generated with microarray technology needs a variety of analytical techniques. Mathematical programming approaches for classification analysis outperform parametric methods when the data depart from assumptions underlying these methods. Therefore, a mathematical programming approach is developed for gene selection and tissue classification using gene expression profiles. RESULTS: A new mixed integer programming model is formulated for this purpose. The mixed integer programming model simultaneously selects genes and constructs a classification model to classify two groups of tissue samples as accurately as possible. Very encouraging results were obtained with two data sets from the literature as examples. These results show that the mathematical programming approach can rival or outperform traditional classification methods.

In this paper we address the problem of identifying all nondominated criterion vectors in large, finite sets of criterion vectors. Two methods, quad-trees and linear lists, are studied in detail. In discussing the methods, a complete algorithmic description of the quad-tree approach, which builds upon the description given by Habenicht (W. Habenicht, ENUQUAD, an interactive DSS-tool for discrete vector optimization problems, M. Cerny, D. Glückaufova, D. Loula, eds. Multicriteria Decision Making: Methods–Algorithms–Applications, Institute of Economics, Prague, pp. 66–81, 1992.), is specified and computational results are reported. The results show that the quad-tree approach is faster than the linear list approach on problems with large sets of criterion vectors that have more than 2, but fewer than 8, objectives. Also, in general, the larger the proportion of vectors that are nondominated in the set of criterion vectors, the more effective the quad-tree approach.