Annals of Operations Research最新文献

We consider a shortest-path network interdiction problem under endogenous uncertainty on successful detection. Endogenous uncertainty arises from the fact that the interdictor may decide to enforce surveillance on some critical arcs, which would affect the prior probability of success on those arcs. The evader decision is formulated as a two-stage stochastic programming problem. In a “here and now situation”, he has to choose the shortest path in the network before realizing detection scenarios. Then, in the second stage, the evader tries to minimize the expected cost of being detected over all possible scenarios. We consider binary scenarios to represent whether or not the evader is detected on each path and apply a linearization method to deal with the non-linearity in the decision-dependent probability measure. A decomposition method is used to solve the proposed model for a case study of a worldwide drug trafficking network. The case study is concerned with finding the most critical arcs for interdicting drug trafficking. Numerical results show that a tiny increase in the probability of opium seizures leads to a significant change in the expected cost when the critical arcs are interdicted. Due to the exponential number of scenarios, the model could not be solved in a reasonable time. Common scenario reduction methods are designed for exogenous uncertainty. We apply an improved bundling method to reduce the number of scenarios in case of endogenous uncertainty. Computational results show that our method reduces the model size and solution time tremendously without significantly affecting the objective value.

Outsourcing operations have become an essential factor in enhancing the competitive advantage of software development enterprises. In this work, we examine the application of combinatorial auction in technician assignment and outsourcing service procurement, which is conducted by software enterprises to minimize the total cost of developing all the software. It gives rise to an unrelated parallel machine scheduling problem incorporating combinatorial auction (UPMSCA). Here, the jobs represent the software to be developed, and they consume the perishable time resources of the development technicians, which can be translated into monetary costs. The objective is to schedule the jobs on parallel machines or select the bid with the lowest cost. To solve the problem, we propose an arc-flow model and a set-partitioning formulation with column-based constraints. A branch-and-price algorithm with four branching rules is proposed and utilizes an effective dynamic programming algorithm to solve the pricing subproblem in the pattern-based formulation. To speed up computation, a bidirectional search method and a dominance rule are applied. Results from extensive computational tests on 100 sets of randomly generated instances demonstrate the performance of our algorithm.

The systematic prediction of real estate prices is a foundational block in the operations of many firms and has individual, societal and policy implications. In the past, a vast amount of works have used common statistical models such as ordinary least squares or machine learning approaches. While these approaches yield good predictive accuracy, most models work very differently from the human intuition in understanding real estate prices. Usually, humans apply a criterion known as “direct comparison”, whereby the property to be valued is explicitly compared with similar properties. This trait is frequently ignored when applying machine learning to real estate valuation. In this article, we propose a model based on a methodology called prototype-based learning, that to our knowledge has never been applied to real estate valuation. The model has four crucial characteristics: (a) it is able to capture non-linear relations between price and the input variables, (b) it is a parametric model able to optimize any loss function of interest, (c) it has some degree of explainability, and, more importantly, (d) it encodes the notion of direct comparison. None of the past approaches for real estate prediction comply with these four characteristics simultaneously. The experimental validation indicates that, in terms of predictive accuracy, the proposed model is better or on par to other machine learning based approaches. An interesting advantage of this method is the ability to summarize a dataset of real estate prices into a few “prototypes”, a set of the most representative properties.

Sports timetabling is a complex and challenging problem. The latest open benchmark dataset for the sport timetabling problem is from the International Timetabling Competition (ITC) 2021. Due to its complexity, only a few approaches have successfully generated feasible solutions for the problems in this dataset, as reported in scientific literature. To the best of our knowledge, there is only one study in the literature that has successfully generated feasible solutions for all 45 problems in the dataset. In this paper, we propose our novel efficient algorithm based on the Iterated Local Search algorithm to solve the ITC 2021 benchmark dataset. Unlike prior successful approaches that combined metaheuristics with an exact approach, our proposed approach is solely metaheuristic. Our contribution includes the design of strategies for both perturbation and local search phases, coupled with the integration of shuffling strategies. The experimental results show that our proposed algorithm is remarkably successful in generating feasible solutions for all 45 problems present in the ITC 2021 dataset.

Volatility in the supply chain of critical products, notably the vaccine shortage during the pandemic, influences livelihoods and may lead to significant delays and long waiting times. Considering the capital- and time-intensive nature of capacity expansion plans, strategic operational production decisions are required best to address the supply-demand mismatches given the limited production resources. This research article investigates the production scenarios where the demand of one agent must be completed within a specified due date, hereinafter referred to as the deadline, while minimizing the maximum or total completion time of another agent's demand. For this purpose, the Two-Agent Proportionate Flowshop Scheduling Problem with deadlines is introduced. Two polynomial-time optimization algorithms are developed to solve these optimization problems. This study will serve as a basis for further developing this practical yet understudied scheduling problem.

Linear programming has had a tremendous impact in the modeling and solution of a great diversity of applied problems, especially in the efficient allocation of resources. As a result, this methodology forms the backbone of introductory courses in operations research. What students, and others, may not appreciate is that linear programming transcends its linear nomenclature and can be applied to an even wider range of important practical problems. The objective of this article is to present a selection, and just a selection, from this range of problems that at first blush do not seem amenable to linear programming formulation. The exposition focuses on the most basic models in these selected applications, with pointers to more elaborate formulations and extensions. Thus, our intent is to expand the modeling awareness of those first encountering linear programming. In addition, we hope this article will be of interest to those who teach linear programming and to seasoned academics and practitioners, alike.

Mean–variance optimization, introduced by Markowitz, is a foundational theory and methodology in finance and optimization, significantly influencing investment management practices. This study enhances mean–variance optimization by integrating machine learning-based anomaly detection, specifically using GANs (generative adversarial networks), to identify anomaly levels in the stock market. We demonstrate the utility of GANs in detecting market anomalies and incorporating this information into portfolio optimization using robust methods such as shrinkage estimators and the Gerber statistic. Empirical analysis confirms that portfolios optimized with anomaly scores outperform those using conventional portfolio optimization. This study highlights the potential of advanced data-driven techniques to improve risk management and portfolio performance.