Dynamic Bottleneck in Serial Production Lines with Bernoulli Machines

A manufacturing process is typically modeled as a stochastic process in production systems research. A challenge commonly met in practical applications is the identification of the parameters for such stochastic process models. The conventional approach focuses on collecting data from each individual work-station’s operation and relies on the modeler’s training and experience to convert the raw data into model parameters. A new modeling approach has recently been proposed that uses system performance metrics (e.g., throughput, work-in-process) as inputs to reversely calculate the model parameters. In this paper, we investigate the efficacy of statistical learning methods in solving the problem of production system model parameter identification. Specifically, three commonly used methods, multivariate regression, random forest, and artificial neural network, are applied to parameter identification in Bernoulli serial line models. Numerical experiments demonstrate the performance of these methods and show that the machine parameters can be identified with high accuracy.

Two analytical models for evaluating performance of Gigabit Ethernet hosts with finite buffer

Summary of United States Human Proteome Organisation (HUPO) Symposium Entitled “Standardized Clinical Proteomics Platforms”

Adaptive Body Biasing (ABB) is a popularly used technique to mitigate the increasing impact of manufacturing process variations on leakage power dissipation. The efficacy of the ABB technique can be improved by partitioning a design into a number of islands, each with its individual body-bias voltage. In this paper, we propose a system-level leakage variability mitigation framework to partition a multiprocessor system into body-bias islands at the processing element (PE) granularity at design time, and to optimally assign body-bias voltages to each island post-fabrication. As opposed to prior gate- and circuit-level partitioning techniques that constrain the global clock frequency of the system, we allow each island to run at a different speed and constrain only the relevant system performance metrics - in our case the execution deadlines. Experimental results show the efficacy of the proposed framework in reducing the mean and standard deviation of leakage power dissipation compared to a baseline system without ABB. At the same time, the proposed techniques provide significant runtime improvements over a previously proposed Monte-Carlo based technique while providing similar reductions in leakage power dissipation.

Optimizing directly the system performance metric is an important method for correcting wave-front distortions in adaptive optics (AO) systems. In extended object imaging, some classical image quality metrics are often used as system performance evaluation function, which is the optimized object of the control algorithm. But those metrics do not consider the existence of imaging noise. Practically, the observed object images are degraded not only by the atmospheric turbulence but also imaging system noise. The noise in image will affect the value of image quality criteria and further affect the correction capability of AO system. An AO system with Stochastic Parallel Gradient Descent (SPGD) algorithm and a 61-element deformable mirror is simulated to restore the image of a turbulence-degraded extended object and the gray level variance function acted the optimized object by control algorithm. The analysis results show that the correction capability is affected severely when the turbulence strength is relatively big (D/r0<10). This paper presents a method that the observed image is processed by the bilateral filter before the system performance metric is calculated, so that the effect of noise on the system performance metric can be mitigated. Numerical simulation results verify the method is effective and the image quality based on AO technique and the bilateral filter is much better than that of only using AO technique.

Model-based analysis of production systems is one of the main areas in manufacturing research. The foundation of the successful application of these theoretical studies is the availability of valid and high-fidelity mathematical models that are capable of capturing the behavior of job flow in production systems. The modeling process of a production system, however, may require a significant amount of nonstandardized work that can only be done properly by someone with solid training in the area and extensive experience through real case studies. This poses a critical challenge in the effective implementation of these valuable theoretical results in the Industry 4.0 era. To overcome this, we propose a new production systems modeling paradigm inspired by system identification: calculate production system model parameters that best match the standard system performance metrics measured on the factory floor. Specifically, in this article, we consider production lines characterized by the Bernoulli serial line model and develop algorithms that identify model parameters to fit the system throughput and work-in-process. Analytical algorithms are derived to solve this problem in a two-machine line case and then extended to multi-machine lines. The accuracy and computational efficiency of the algorithms are demonstrated through extensive numerical experiments. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —A high-fidelity mathematical model is of critical importance to the implementation of any model-based production system analysis method. Currently, the construction of such models is carried out in an ad hoc manner. The quality of the resulting models may heavily depend on the training, experience, intuition, and personal preference of the modeler. The proposed model parameter identification method focuses on standard key performance indices commonly measured on the factory floor. The advantage is twofold. First, these standard performance metrics are consistently defined regardless of industry, thus avoiding any data-ambiguity issue that may occur when using complex machine/equipment status data. Second, measuring these performance metrics in real time is typically convenient and cost effective, even for manufacturing plants without high-end IT infrastructure, thus making the technology accessible to not only large but also small- and mid-sized manufacturers. Using the algorithms developed in this article, a practitioner can quickly construct a serial production line model and then utilize it to access the rich library of production analysis, design, and control methods available in the literature.

We distinguish two types of metric for the evaluation of rule-based learning systems: performance metrics are derived from the feedback to the learning agent from its teacher or environment, while population state metrics are derived from inspection of the rule base used for decision making. We propose novel population state metrics for use with learning classifier systems, evaluate them using the XCS system, and demonstrate their superiority in some cases.

The literature on queueing systems with finite buffers addresses mostly asymptotic performance metrics on an aggregate flow, and/or generally relies on a convenient, but provably inaccurate, approximation of the loss probability by the overflow probability in an infinite size buffer. This paper addresses non-asymptotic per-flow metrics in a multi-flow queueing system with finite buffer and FIFO scheduling. The analysis dispenses with the above approximation, and lends itself to several interesting insights on the impact of finite buffers on per-flow metrics. Counterintuitively, the per-flow delay distribution is not monotonous in the buffer size, and such an effect is especially visible in high burstiness regimes. Another observation is that buffer dimensioning becomes insensitive to the type of SLA constraint, e.g., fixed violation probability on either loss or delay, in high multiplexing regimes. In the particular case of aggregate scheduling, the results on the aggregate input flow significantly improve upon existing results by capturing the manifestation of bufferless multiplexing in regimes with many flows.

Cloud computing alters the way organizations manage and deploy their IT resource. It provides an organization with scalable, inexpensive, and flexible options. The complexity and dynamic nature of cloud environments pose a challenge to maintaining high availability at all times, especially when the system fails or a disaster arises. The legacy techniques of disaster recovery, fault tolerance, and high availability leave behind much to be desired. These techniques are mostly static, slow to respond, and have a dismal ability to adapt to continuously changing conditions in contemporary cloud systems. Such techniques largely depend on manual configurations and predefined policies; resulting in lots of inefficiencies and increases in the risk of service downtime. This research investigates the way Artificial Intelligence (AI) changes the paradigm on cloud resilience to promote adoption of intelligent systems for guaranteed disaster recovery, better fault-tolerant behavior, or transparent high availability. With machine learning algorithms, AI-based cloud services utilize large data volumes to reveal patterns within system logs, performance metrics, and user behavior data; thereby offering real-time anomaly detection and predictive failure analysis. For example, techniques like predictive analytics help cloud providers predict likely system outages, optimize the resources to be used, and automate failover processes (Xu et al., 2021; Lee &amp; Kumar, 2022). AI-aided disaster recovery techniques employ complex algorithms to produce an adaptive backup mechanism, thereby minimizing loss and reducing restoration time. Fault tolerance in AI cloud systems comes from intelligent error correction, automatic isolation of faults, and self-healing features, i.e. repair of faults without the need for human supervision (Chen et al., 2020). Besides that, AI also contributes to high availability through intelligent load balancing, which ensures that at any given time, resources are optimally distributed throughout the network to sustain continuous service even during peak demand or unanticipated failures (Patel &amp; Zhang, 2023). The approach is a comprehensive review of the various existing literature on the topic, empirical analysis of the current AI-driven cloud solutions available in the market, and case studies for comparison analysis on the different AI systems. The study scenario reveals that AI-driven solutions noticeably reduce downtimes, lead to improved recovery times, and contribute to overall system reliability as compared to traditional methods. However, other areas include model bias, data privacy, and continuous training of AI models. This study expands the trends of AI in the field of cloud computing by documenting the significance of intelligent systems in bridging traditional weaknesses of resilience strategies. It further signifies the need for AI into predictive maintenance, automated disaster response, and proactive fault management of rapidly changing dynamic cloud environments. Future studies will focus on AI integration along with edge computing and blockchain technologies for even more robust and secure services in cloud operations.

An M/G/c/c state dependent network is a quantitative model for replicating and analyzing the behavior of occupants travelling through a network. The model, however, assumes that its inter-arrival times are independent and exponentially distributed, which inadequately reflects the dynamic and high-stakes nature in emergency situations. This paper simulates the M/G/c/c model for emergency evacuations using Erlang-k distributions—representing the real-world arrival patterns of evacuees in a more controlled and less random arrivals—and correlated inter-arrival times—common during evacuations as occupants moving in clusters due to panic or structural flow patterns. To achieve this, an M/G/c/c simulation model was first developed using Arena simulation software. The model was used to analyze the impact of various arrival rates on system performance metrics. Consequently, the performance metrics were compared with those obtained by replacing exponential inter-arrival times with Erlang-k distributions and independent arrivals with correlated cluster arrivals. The results show that Erlang-k distributions lead to better performance and smoother flow since the arrival is more controlled, while correlated arrivals increase congestion.

Mobile data traffic is currently growing exponentially and these rapid increases have caused the backhaul data rate requirements to become the major bottleneck to reducing costs and raising revenue for operators. To address this problem, caching techniques have attracted significant attention since they can effectively reduce the backhaul traffic by eliminating duplicate data transmission that carries popular content. In addition, other system performance metrics can also be improved through caching techniques, e.g., spectrum efficiency, energy efficiency, and transmission delay. In this paper, we provide a systematical survey of the state-of-the-art caching techniques that were recently developed in cellular networks, including macro-cellular networks, heterogeneous networks, device-to-device networks, cloud-radio access networks, and fog-radio access networks. In particular, we give a tutorial on the fundamental caching techniques and introduce caching algorithms from three aspects, i.e., content placement, content delivery, and joint placement and delivery. We provide comprehensive comparisons among different algorithms in terms of different performance metrics, including throughput, backhaul cost, power consumption, and network delay. Finally, we summarize the main research achievements in different networks, and highlight main challenges and potential research directions.

Uncertainty propagation is an important step in technology forecasting. Uncertainty in technology impacts and other system variables must be propagated into the system performance metrics that will be used to judge competing technology portfolios. Monte Carlo simulation and Evidence Theory are two uncertainty propagation techniques commonly used for this purpose. This paper explores the differences between these two techniques when applied to a representative aircraft engine technology forecasting problem. As an example problem, uncertainties in compressor, turbine, and combustor efficiencies are propagated into seven aircraft engine performance metrics. A surrogate model built from NPSS and WATE++ data provides modeling and simulation of the system. A representative set of expert opinion derived uncertainty intervals drives this process. The MATLAB IPP Toolbox is used to conduct Evidence Theory uncertainty propagation on the example problem. The results presented show the impact of sample size selection (number of cases), method selection (Monte Carlo or Evidence Theory), and mixing rule (Dempster’s rule or weighted mixing) on the resulting CDFs. The increase in runtime with sample size for both methods is shown, as well.

An increasing concern when dealing with critical systems or services design is their preparation for a wide range of potentially uncertain operating conditions. In this chapter a four-step simulation-driven decision making methodology (SDDM) is presented to address sensitivity analysis of candidate designs to uncertainty and extreme operating conditions. The approach accounts for, not only the conventional average system performance metrics, but also (i) upper-tail or extreme values of these metrics, and (ii) performance sensitivity measures to uncertainties in the simulation model. The example used to illustrate the application of this technique is a hospital emergency department design case study wherein different design alternatives are compared using patient time-in-system performance metric under multiple uncertain operating conditions.

A novel approach for low mutual coupling and ultra-compact Two Port MIMO antenna development for UWB wireless application

HTTP video streaming employs the TCP protocol as its transport protocol, which provides reliability on one hand, but delay or delay jitter on the other hand. IEEE 802.11 MAC protocol is assumed to support wireless networks, whose channel access efficiency impacts the performance of the upper layer, i.e., the data rate and the delay of TCP. In this paper, an analytical model for Quality of Experience (QoE) of HTTP video streaming over single-hop wireless networks is proposed. Particularly, two QoE performance metrics are evaluated, which are underflow times and fluency of playout, defined as the buffer empty times during playout, and the ratio of ideal watching time to the real watching time. The impacts of system parameters, such as the length of a video file, the playout threshold, the playout rate, and the arriving rate of a TCP streaming, on the system performance metrics are analyzed in detail. It is found that when both the TCP congestion window and the MAC contention window are set to one, the arriving rate of a TCP streaming can achieve its maximum, which results in the optimal QoE performance. Moreover, the optimal setting for the MAC contention window is also obtained for the case that the TCP congestion window is larger than one. The accuracy of the proposed analytical model is verified by extensive simulations.