Article
Computer Science, Interdisciplinary Applications
Xiaofan Lai, Xiaolong Lu, Xinyao Yu, Ning Zhu
Summary: This study introduces a new vaccination station location model that takes into account the planning of medical professionals, vaccine procurement, and inventory decisions. A two-stage stochastic mixed integer linear program is used to address the uncertain demands for multiple types of vaccines over multiple periods. By developing a heuristic algorithm based on Benders decomposition, the effectiveness and efficiency of the model and new heuristics are demonstrated through numerical experiments and sensitivity analysis.
COMPUTERS & INDUSTRIAL ENGINEERING
(2021)
Article
Engineering, Chemical
Ariel A. Boucheikhchoukh, Christopher L. E. Swartz, Eric Bouveresse, Pierre Lutran, Anna Robert
Summary: Uncertainty in refinery planning poses challenges to the day-to-day operations of an oil refinery. Stochastic programming framework can incorporate parameter uncertainty and provide robust solutions, which is more effective than deterministic modeling techniques.
Article
Management
Jesus A. Rodriguez, Miguel F. Anjos, Pascal Cote, Guy Desaulniers
Summary: The maintenance scheduling problem for hydroelectric generators involves uncertainty in water flows and nonlinearity in hydroelectric production, solved using a two-stage stochastic program and parallelized Benders decomposition algorithm. By approximating hydroelectric production with linear inequalities and indicator variables, tailoring and testing various acceleration techniques successfully sped up the algorithm fourfold. Industrial results confirm high scalability of the parallelized Benders implementation in various scenarios.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2021)
Article
Management
Niels van der Laan, Ward Romeijnders
Summary: We propose a new solution method for two-stage mixed-integer recourse models that can handle general mixed-integer variables in both stages. Our method is based on Benders' decomposition, where we iteratively construct tighter approximations of the expected second stage cost function using a new family of optimality cuts derived from extended formulations of the second stage problems. We show convergence of our method by proving that the optimality cuts recover the convex envelope of the expected second stage cost function. Finally, we demonstrate the potential of our approach through numerical experiments on investment planning and capacity expansion problems.
OPERATIONS RESEARCH
(2023)
Article
Computer Science, Artificial Intelligence
Jinyan Hu, Yanping Jiang
Summary: This paper examines the allocation problem of shared parking spaces considering parking unpunctuality and no-shows. It proposes an effective approach using sample average approximation (SAA) combined with an accelerating Benders decomposition (ABD) algorithm to solve the problem. The numerical experiments demonstrate the significance of supply-demand balance for the operation and user satisfaction of the shared parking system.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Management
Xavier Blanchot, Francois Clautiaux, Boris Detienne, Aurelien Froger, Manuel Ruiz
Summary: This paper presents a new exact algorithm for solving two-stage stochastic linear programs. The algorithm, based on the multicut Benders reformulation, divides the subproblems into batches and solves only a small proportion of them in each iteration. A general framework is proposed to stabilize the algorithm, and its finite convergence and exact behavior are demonstrated. Computational experiments on large-scale stochastic optimization instances show the efficiency of the proposed algorithm compared to nine alternative algorithms in the literature. Additional computational time savings are obtained using primal stabilization schemes.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2023)
Article
Management
Mahboobeh Peymankar, Morteza Davari, Mohammad Ranjbar
Summary: This paper discusses the maximization of expected net present value of a project under uncertain cash flows using discrete scenarios. It proposes two ILP formulations and two-stage stochastic programming approaches, utilizing Benders decomposition, to address the problem efficiently. The computational results demonstrate that the developed Benders-based methods outperform the ILP formulations.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2021)
Article
Computer Science, Software Engineering
Giacomo Nannicini, Emiliano Traversi, Roberto Wolfler Calvo
Summary: The paper proposes a nested decomposition scheme for infinite-horizon stochastic linear programs, ensuring convergence to a certain confidence level by exploring finite-horizon problems. The algorithm shows high accuracy in solving instances under random generation.
MATHEMATICAL PROGRAMMING COMPUTATION
(2021)
Article
Economics
Santiago Nieto-Isaza, Pirmin Fontaine, Stefan Minner
Summary: This study proposes a crowd-shipping solution supported by strategically located mini-depots. It addresses the network design problem and conducts computational experiments. The results demonstrate the potential for utilizing stochastic crowd flows to deliver packages and achieving cost savings.
TRANSPORTATION RESEARCH PART B-METHODOLOGICAL
(2022)
Article
Computer Science, Software Engineering
Leonardo Lozano, J. Cole Smith
Summary: This article discusses a special class of two-stage stochastic integer programming problems and proposes an approach to finding strong dual formulations using dynamic programming methods. The efficacy of the approach is demonstrated on stochastic traveling salesman problems.
MATHEMATICAL PROGRAMMING
(2022)
Article
Management
Sheng Liu, Zhixing Luo
Summary: This study proposes a novel structured approximation framework for the stochastic dynamic driver dispatching and routing problem in last-mile delivery systems, focusing on on-time performance. The framework approximates the value function and establishes its performance guarantee under large-demand scenarios. Efficient exact algorithms based on Benders decomposition and column generation are developed to provide verifiably optimal solutions within minutes. Evaluation on real-world data shows that our framework outperforms the current company policy by 36.53% on average in terms of delivery time. Several policy experiments with varying fleet sizes and dispatch frequencies are conducted to understand the value of dynamic dispatching and routing.
M&SOM-MANUFACTURING & SERVICE OPERATIONS MANAGEMENT
(2023)
Article
Computer Science, Interdisciplinary Applications
Bart Smeulders, Valentin Bartier, Yves Crama, Frits C. R. Spieksma
Summary: This paper addresses the issue of optimizing transplant matchings in kidney exchange programs, specifically focusing on the uncertainty in compatibility assessment between donors and recipients. While previous literature has explored this issue, this paper explicitly formulates the maximization of expected transplants as a two-stage stochastic integer programming problem, revealing computational challenges. Various algorithmic approaches are proposed and tested, leading to kidney exchanges of higher quality compared to earlier models.
INFORMS JOURNAL ON COMPUTING
(2022)
Article
Energy & Fuels
Fernando Garcia-Munoz, Sebastian Davila, Franco Quezada
Summary: New consumer-centric electricity market schemes are emerging as an alternative to manage energy balancing in distribution networks with high distributed energy resources penetration. Peer-to-peer energy trading is an attractive scheme for reducing user and distribution system operator costs by exchanging energy surplus between energy community users. This article presents a two-stage stochastic mixed-integer linear programming model to address the day-ahead scheduling problem in an energy community operating under a peer-to-peer energy trading scheme. The proposed model minimizes community costs, considers network limitations, and allows consumers to act as buyers or sellers depending on their consumption and self-generation.
Article
Computer Science, Interdisciplinary Applications
Nader Ghaffarinasab
Summary: This paper introduces the hub location problem with Bernoulli demands and presents a two-stage stochastic programming model under single and multiple allocation settings. To solve large instances, exact solution algorithms based on Benders decomposition and Lagrangian relaxation are proposed. Extensive computational experiments are conducted to test the efficiency of the proposed models and algorithms, as well as to evaluate the effect of different input parameters on the optimal solutions.
COMPUTERS & OPERATIONS RESEARCH
(2022)
Article
Engineering, Chemical
Ilias Mitrai, Wentao Tang, Prodromos Daoutidis
Summary: This article proposes using stochastic blockmodeling (SBM) to learn the underlying block structure in generic optimization problems and estimates the interconnection patterns through parametric statistical inference for decomposition-based solution algorithms. Furthermore, a general software platform is developed for automated block structure detection and following distributed and hierarchical optimization approaches.
Article
Computer Science, Hardware & Architecture
Yang Zhang, Zehui Xiong, Dusit Niyato, Ping Wang, Zhu Han
Article
Engineering, Electrical & Electronic
Yang Zhang, Zehui Xiong, Dusit Niyato, Ping Wang, H. Vincent Poor, Dong In Kim
IEEE TRANSACTIONS ON COMMUNICATIONS
(2020)
Article
Computer Science, Information Systems
Amirheckmat Taghizadeh, Hamed Kebriaei, Dusit Niyato
IEEE INTERNET OF THINGS JOURNAL
(2020)
Article
Computer Science, Information Systems
Cheng Zhan, Han Hu, Xiufeng Sui, Zhi Liu, Dusit Niyato
IEEE INTERNET OF THINGS JOURNAL
(2020)
Article
Computer Science, Information Systems
Yi Liu, James J. Q. Yu, Jiawen Kang, Dusit Niyato, Shuyu Zhang
IEEE INTERNET OF THINGS JOURNAL
(2020)
Article
Computer Science, Information Systems
Yanru Zhang, Dusit Niyato, Ping Wang, Zhu Han
IEEE INTERNET OF THINGS JOURNAL
(2020)
Article
Computer Science, Information Systems
Joohyung Lee, DaeJin Kim, Dusit Niyato
IEEE INTERNET OF THINGS JOURNAL
(2020)
Article
Computer Science, Information Systems
Bin Lyu, Dinh Thai Hoang, Shimin Gong, Dusit Niyato, Dong In Kim
IEEE WIRELESS COMMUNICATIONS LETTERS
(2020)
Article
Engineering, Electrical & Electronic
Helin Yang, Zehui Xiong, Jun Zhao, Dusit Niyato, Qingqing Wu, H. Vincent Poor, Massimo Tornatore
Summary: This study focuses on utilizing intelligent reflecting surfaces to mitigate malicious interference caused by smart jammers and enhance communication performance by optimizing power allocation and reflecting beamforming. The proposed fuzzy win or learn fast-policy hill-climbing (WoLF-CPHC) learning method efficiently improves the effectiveness of anti-jamming power allocation and reflecting beamforming strategies.
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS
(2021)
Article
Telecommunications
Zehui Xiong, Yang Zhang, Wei Yang Bryan Lim, Jiawen Kang, Dusit Niyato, Cyril Leung, Chunyan Miao
Summary: This paper proposes a Markov decision process (MDP) model to optimize energy and data transfer for UAV-assisted communication in massive machine type communication (mMTC). The MDP model is solved by value iteration algorithm to obtain optimal strategies for the UAV, and further extended to address uncertainties, partially observable states, and large state space. Simulation results demonstrate that the MDP model with deep reinforcement learning (DRL) can achieve better wireless energy and data transfer strategies compared to baseline schemes.
IEEE TRANSACTIONS ON COGNITIVE COMMUNICATIONS AND NETWORKING
(2021)
Article
Engineering, Civil
Jer Shyuan Ng, Wei Yang Bryan Lim, Hong-Ning Dai, Zehui Xiong, Jianqiang Huang, Dusit Niyato, Xian-Sheng Hua, Cyril Leung, Chunyan Miao
Summary: This paper proposes the use of UAVs as wireless relays to improve the accuracy of FL by facilitating communication between IoV components and the FL server, and presents a joint auction-coalition formation framework to address the allocation of UAV coalitions, maximizing individual profits.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2021)
Article
Computer Science, Cybernetics
Chia-Yu Lin, Yu-Fang Chiu, Li-Chun Wang, Dusit Niyato
Summary: A multilayer LTM (ML-LTM) was introduced to handle the hierarchical clustering issues of multicontent variables and to develop a multicontent recommendation system. Experimental results showed that ML-LTM achieved a higher recommendation accuracy and an incremental update approach was proposed to save updating time.
IEEE TRANSACTIONS ON COMPUTATIONAL SOCIAL SYSTEMS
(2021)
Article
Engineering, Electrical & Electronic
Sahar Idrees, Xiangyun Zhou, Salman Durrani, Dusit Niyato
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS
(2020)
Proceedings Paper
Computer Science, Information Systems
Zehui Xiong, Wei Yang Bryan Lim, Jiawen Kang, Dusit Niyato, Ping Wang, Chunyan Miao
2020 IEEE WIRELESS COMMUNICATIONS AND NETWORKING CONFERENCE (WCNC)
(2020)
Article
Computer Science, Hardware & Architecture
Junjie Tan, Ying-Chang Liang, Nguyen Cong Luong, Dusit Niyato
Summary: Traditional centralized learning networks are facing challenges in privacy preservation and communication overheads, leading to the proposal of federated learning networks as a promising alternative. FLNs utilize edge devices for distributed training, preserving privacy and reducing communication overheads. However, the reliance on contributions from all devices can make the training process vulnerable to poisoning attacks.
