Article
Management
Btissam Er-Rahmadi, Tiejun Ma
Summary: In this paper, a new Mixed Integer Linear Programming (MILP) optimization-based failure detector (FD) algorithm is proposed. The MILP formulation is obtained via piecewise linearization relaxations and aims to find optimal FD parameters that meet the desired system requirements. The proposed approach improves overall FD performance and scalability by considering network conditions and system parameters as constraints and adapting to real-time network changes. The results obtained from testing in a realistic environment show consistent improvement in the performance and scalability of the FD. This paper is the first attempt to combine MILP-based optimization modeling with FD to achieve performance guarantees.
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH
(2022)
Article
Automation & Control Systems
Antoine Dedieu, Hussein Hazimeh, Rahul Mazumder
Summary: This study utilizes mixed-integer programming to solve l(0)-regularized regression problems, introduces new scalable algorithms, and presents new estimation error bounds. Experimental results demonstrate significant improvements in statistical performance compared to competing methods.
JOURNAL OF MACHINE LEARNING RESEARCH
(2021)
Article
Thermodynamics
Yingzong Liang, Chi Wai Hui, Xianglong Luo, Jianyong Chen, Zhi Yang, Ying Chen
Summary: The proposed equation-based optimization framework provides a comprehensive modeling structure for various thermodynamic cycle design problems, offering efficient and effective solutions for solving thermodynamic cycle and related energy systems optimization problems.
ENERGY CONVERSION AND MANAGEMENT
(2021)
Article
Computer Science, Interdisciplinary Applications
Victor Blanco, Ricardo Gazquez
Summary: This paper analyzes a continuous version of the maximal covering location problem and proposes a mathematical programming framework as well as different resolution strategies, including two branch-and-cut approaches and a math-heuristic algorithm. Results of extensive computational experiments comparing performance are reported.
COMPUTERS & OPERATIONS RESEARCH
(2021)
Article
Biochemical Research Methods
Enio Gjerga, Aurelien Dugourd, Luis Tobalina, Abel Sousa, Julio Saez-Rodriguez
Summary: Protein post-translational modifications are crucial for cellular processes, and mass spectrometry analysis of proteome modifications can provide insights into signaling mechanisms. The newly formulated PHONEMeS method as an Integer Linear Program (ILP) is significantly more efficient and can analyze data sets with multiple time points to understand signal propagation dynamics. This study expands the analysis scenarios and sheds light on signaling mechanisms and drug modes of action.
JOURNAL OF PROTEOME RESEARCH
(2021)
Article
Engineering, Chemical
Florian Joseph Baader, Andre Bardow, Manuel Dahmen
Summary: The increasing volatility of electricity prices highlights the importance of simultaneous scheduling optimization for production processes and their energy systems. In this study, we propose an efficient scheduling formulation that takes into account both process dynamics and binary on/off-decisions in the energy system. By considering three different aspects, we demonstrate the feasibility of achieving fast optimization for real-time scheduling.
Article
Mathematics, Applied
Kurt M. Anstreicher
Summary: This method utilizes the computational power of modern MILP solvers to test if a given matrix is copositive by solving a single mixed-integer linear programming problem. Numerical experiments demonstrate that the method is robust and efficient.
LINEAR ALGEBRA AND ITS APPLICATIONS
(2021)
Article
Computer Science, Interdisciplinary Applications
Christopher Muir, Alejandro Toriello
Summary: Motivated by applications in cloud computing, this paper studies interval scheduling problems with economies of scale. Each job has start time, end time, and a cost function for scheduling a subset of jobs on the same machine. The goal is to minimize the total schedule cost by partitioning the jobs, considering overlapping jobs cannot be processed on the same machine. The paper proposes a set cover formulation and a column generation algorithm to solve the linear relaxation. The methods are computationally evaluated and shown to outperform known integer programming formulations.
COMPUTERS & OPERATIONS RESEARCH
(2023)
Article
Automation & Control Systems
Rubens J. M. Afonso, Roberto K. H. Galvao
Summary: This note addresses the problem of crossing a target set between sample instants under the influence of bounded unknown disturbances. The proposed solution utilizes mixed-integer linear programming and is less conservative compared to the standard approach of imposing pointwise-in-time constraints at the sample instants.
INTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROL
(2021)
Review
Biochemistry & Molecular Biology
Pravin D. Patil, Sakshi Salokhe, Aparna Karvekar, Prabhavati Suryavanshi, Ajay N. Phirke, Manishkumar S. Tiwari, Shamraja S. Nadar
Summary: Conventional techniques for enzyme immobilization have limitations in terms of activity recovery, stability, and applicability, while microfluidic devices offer precise control over immobilization parameters and result in highly active immobilized enzymes. This review comprehensively explains the design, construction, and methods of enzyme immobilization using microfluidic devices, highlighting the advantages and applications. It also discusses future developments and challenges in this field.
INTERNATIONAL JOURNAL OF BIOLOGICAL MACROMOLECULES
(2023)
Article
Thermodynamics
Karl Vilen, Erik O. Ahlgren
Summary: Most computer models used in energy systems optimization modeling studies are constructed using linear equations. However, linear equations may not adequately reflect real-world conditions and are less suitable for representing individual-scale technologies in local system studies. This study investigates the differences in heating solutions and model solution times for a local expanding heating system. The results show that the use of district heating is higher for cost structures that use mixed integer linear programming. On the other hand, the solution time is significantly shorter for linear formulations compared to mixed integer linear formulations.
Article
Green & Sustainable Science & Technology
Xinyu Wu, Yiyang Wu, Xilong Cheng, Chuntian Cheng, Zehong Li, Yongqi Wu
Summary: Optimizing hydro unit commitment has the potential to improve water use efficiency, but it is necessary to consider complex constraints from power grid, hydropower station, and unit operation. To overcome the problem of conflicting constraints leading to no feasible solution, a constraint grading principle is proposed to convert hard constraints into soft constraints and rank them in priority levels. The proposed method effectively solves the problem of no feasible solution due to conflicting constraints in HUC.
Article
Management
Mercedes Pelegrin, Liding Xu
Summary: This paper explores the continuous set-covering location problem on a network and proposes a Mixed Integer Linear Programming formulation. The model is enhanced by a preprocessing algorithm and valid inequalities. Additionally, a scalable approach is introduced that is suitable for real-world networks. Computational experiments are conducted to test the proposed method against an existing exact method.
OMEGA-INTERNATIONAL JOURNAL OF MANAGEMENT SCIENCE
(2023)
Article
Computer Science, Hardware & Architecture
Brian Crites, Cody Falzone, Tristan Lopez, Karen Kong, Philip Brisk
Summary: Seam carving is an algorithm for image size reduction that removes least important paths from images iteratively. This article applies seam carving to flow-based microfluidic chips and presents three variants, finding that nonrectilinear carving yields the best results.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2021)
Article
Automation & Control Systems
Rubens J. M. Afonso, Roberto K. H. Galvao, Gabriel A. Souza, Marcos R. O. A. Maximo, Angelo Caregnato-Neto
Summary: This note discusses the problem of k-hop connectivity in a network of mobile agents. It proposes linear constraints and binary optimization variables to ensure k-hop connectivity, which are integrated into a mixed-integer linear programming (MILP) trajectory planning model. Simulation results demonstrate the application of the proposed method and the impact of varying k in a mission involving multiple targets.
INTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROL
(2023)
Article
Computer Science, Hardware & Architecture
Felix Last, Ulf Schlichtmann
Summary: Supervised machine learning requires large amounts of labeled data for training. In the context of power, performance, and area estimation for embedded memories, transfer learning can be used to reduce the need for supervised data by pre-training neural networks on related domains. This study shows that exploiting similarities among different memory compilers, versions, and technology nodes can significantly reduce the provisioning times for new compiler versions, speeding up the design cycle. Additionally, a new method called domain equalization is proposed to enable transfer learning across structurally different domains.
ACM TRANSACTIONS ON DESIGN AUTOMATION OF ELECTRONIC SYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Xing Huang, Youlin Pan, Zhen Chen, Wenzhong Guo, Lu Wang, Qingshan Li, Robert Wille, Tsung-Yi Ho, Ulf Schlichtmann
Summary: Due to the complexity of chip architecture and assay protocol, traditional methods for designing continuous-flow microfluidic lab-on-a-chip systems have serious gaps between different steps, leading to design failure. This article proposes a novel "one-pass design paradigm" that integrates all design steps seamlessly, eliminating the gaps and generating optimized biochip architectures without the need for design adjustments and modifications. Experimental results demonstrate the effectiveness of the proposed automation flows.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Tinghuan Chen, Grace Li Zhang, Bei Yu, Bing Li, Ulf Schlichtmann
Summary: This article discusses the potential of machine learning methods in addressing the challenges posed by the increasing complexity and size of design space in integrated circuit (IC) design. It provides a comprehensive survey of the current state of knowledge in both IC design problems and ML-based solutions. The article also summarizes the open problems at the intersection of advanced IC design and ML.
IEEE DESIGN & TEST
(2023)
Article
Computer Science, Hardware & Architecture
Mengchu Li, Yushen Zhang, Ju Young Lee, Hudson Gasvoda, Ismail Emre Araci, Tsun-Ming Tseng, Ulf Schlichtmann
Summary: This study proposes the first built-in-self-test (BIST) method for microfluidic large-scale integration (mLSI) chips, which greatly improves the test efficiency by directly testing the control channels instead of individual valves. The test module has moderate area overhead and the test method is easy to operate, without the need for specialized software or external pressure sensors. Experiments show that this test approach is sensitive enough to detect defects as small as 10 μm in size and located several centimeters away from the test module.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Xunzhao Yin, Yu Qian, Mohsen Imani, Kai Ni, Chao Li, Grace Li Zhang, Bing Li, Ulf Schlichtmann, Cheng Zhuo
Summary: In this article, we propose NOR-type 2FeFET-1T and NAND-type 2FeFET-2T TCAM designs based on ferroelectric FET (FeFET) to improve energy efficiency. We also introduce a hybrid ferroelectric NAND-NOR (HFNN) TCAM design with a segmented architecture to reduce search delay and energy consumption. The evaluation results show that the proposed TCAM designs consume significantly less search energy compared to the conventional 16T CMOS TCAM, and can also save a considerable amount of GPU energy consumption.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Nicolo Bellarmino, Riccardo Cantoro, Martin Huch, Tobias Kilian, Raffaele Martone, Ulf Schlichtmann, Giovanni Squillero
Summary: This article presents a methodology based on active learning to select the best samples for testing microcontrollers. By using a multilabel technique to improve predictive accuracy, the approach reduces the training set size and the time and cost required for training.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Di Gao, Zeyu Yang, Qingrong Huang, Grace Li Zhang, Xunzhao Yin, Bing Li, Ulf Schlichtmann, Cheng Zhuo
Summary: This article proposes a Bayesian inference-based framework, BRoCoM, which connects device nonidealities and algorithmic training together for robust computing on memristor crossbars. The proposed framework incorporates different levels of nonidealities into prior weight distribution, optimizing neural network weights to accommodate uncertainties and minimize inference degradation.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Christian Bluemm, Bo Liu, Bing Li, Talha Rahman, Md Sabbir-Bin Hossain, Maximilian Schaedler, Ulf Schlichtmann, Maxim Kuschnerov, Stefano Calabro
Summary: In this article, the challenges and options for scaling IM/DD transceivers towards 800 Gbps are discussed. The focus is on CWDM4 PAM4 transmission over a target distance of 10 km in O-band. Chromatic dispersion (CD) is identified as the main challenge, which can be mitigated through digital signal processing. The use of neural network equalization and magnitude weight pruning can significantly reduce hardware complexity without performance loss.
JOURNAL OF LIGHTWAVE TECHNOLOGY
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Weiqing Ji, Xingcheng Yao, Hailong Yao, Tsung-Yi Ho, Ulf Schlichtmann, Xia Yin
Summary: In paper-based digital microfluidic biochips (P-DMFBs), printing conductive electrodes and control lines on the same side of the photo paper results in the control interference issue. This paper proposes an effective Self-Obstacle Avoiding Escape Routing (SOAER) method for P-DMFBs based on an enhanced network flow model. Experimental results demonstrate a 49x speedup in runtime compared to the state-of-the-art method, showing the efficiency and effectiveness of the proposed method with a success rate of 100% and significantly decreased runtime.
PROCEEDINGS OF THE GREAT LAKES SYMPOSIUM ON VLSI 2023, GLSVLSI 2023
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Weiqing Ji, Hailong Yao, Tsung-Yi Ho, Ulf Schlichtmann, Xia Yin
Summary: Microfluidic biochips are promising and versatile in automating biochemical protocols. Machine learning models can enhance the efficiency and scalability of concentration prediction for microfluidic mixers, but struggle with multiple input flow rates. This paper proposes a new concentration prediction method based on the graph attention networks (GAT), which efficiently predicts the generated concentration of random microfluidic mixers with multiple input flow rates.
PROCEEDINGS OF THE GREAT LAKES SYMPOSIUM ON VLSI 2023, GLSVLSI 2023
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Xinda Chen, Rongliang Fu, Junying Huang, Huawei Cao, Zhimin Zhang, Xiaochun Ye, Tsung-Yi Ho, Dongrui Fan
Summary: This paper presents JRouter, an RSFQ router that considers the two-layer planar Manhattan routing model and copes with SPL placement and length-matching multi-terminal routing. Experimental results show that JRouter has significant advantages in reducing routing width compared to other routing algorithms.
PROCEEDINGS OF THE GREAT LAKES SYMPOSIUM ON VLSI 2023, GLSVLSI 2023
(2023)